😎│cool-projects

[ashen cairn]

I am trying to analyse market trends and macro situation. Pretty much as the guy I was answering to I guess, but without a trading perspective, but mid-term portfolio management reactive to the markets

[hot ember]

sounds good.
so your goal is to analyze market trends and the macroeconomic situation for mid-term portfolio management?

then I think getting new data quickly and reliably is crucial.
How did you solve this mission?

[ashen cairn]

Free apis! FRED, Alpha Vantage yfinance (it is not an api, I don’t really know how it works, but it does) to pull yahoo finance, … I tried trading economics and some others but they are very limited

[hot ember]

lol. then I guess your project won't be very useful.

[ashen cairn]

Why? How would you improve it?

[zinc sedge]

I'm working on a sports project that uses CV to track movements and then give recommendations for improvements. I am following a project laid out by someone, they use perplexity pro to generate the code. Does this community find perplexity to be good with coding or do you recommend other AIs?

[stone parcel]

The problem isn’t AI — it’s how we use it.
Most companies don’t give AI the whole picture.
Tetrix does. 🔗 https://medium.com/deskree-ai/almost-every-company-is-underusing-ai-heres-why-4d7365632d21

[orchid yacht]

Best ai video agent:
https://www.opus.pro/agent?ref_id=780JQI2M7
Use my link to get the early access faster...

[river kiln]

https://github.com/Atul8007/Limit_Tracker

[stone parcel]

🚀 Tetrix Overview — Giving AI Context Across Your Entire System

What if your AI could actually understand your code, infrastructure, and services — all at once?

With Tetrix, that’s now possible.
By connecting GitHub and AWS, Tetrix brings deep context awareness to your entire engineering stack — letting your AI reason across real repositories, environments, and cloud resources.

🎥 Watch the quick overview video here → https://www.youtube.com/watch?v=syoaxJla_EY

[raw kernel]

What is this? I can just understand the terms but not the concept of entirety.

[dapper nymph]

I'm working with some colleagues and we are trying to establish a newsletter/podcast that will actually keep track of what's going on in the field of AI.
I created a survey to gauge interest and if you're interested I encourage you to fill this out so we know what you're looking for.
https://forms.gle/36Ycf23zPYG8B4WD9

All the survey results are fully open. Anyone can view it online.

The basic idea is that it would cover new product launches, maybe explain some research papers in simple terms, compare AI search engines, stuff like that.

[pine bluff]

Whenever I see “most companies” / “most people” in any AI content immediately stop reading. I wonder if this is just me?

[vapid cliff]

Isn't that just how accurate things are written?

Most / Probably / Possibly, etc

[pine bluff]

Not quite, it is more the inference that they, and they alone know how things should be done. They are the bearer of all wisdom, and everybody else knows nothing. It is a subtle manipulation that says that you’re most people, unless you submit to their superiority and concede their point - I guess most people just don’t spot the trick - (P.S. there is a touch of hyperbole here)

[vapid cliff]

Or it just means what is says, which is that most people do x.

Most people have breakfast...
I don't think that sentence denoted any superiority

[pseudo vine]

when even your discord msg is generated by AI, sure

[pseudo vine]

you're not alone

[pine bluff]

Somebody should work on getting AI to write click bait articles that don’t come with all of the click bait tropes - “and here’s why” - “and this is what happened” - now that would be a project that would be worth paying for

let’s get AI to not write the same slop that humans have been pumping out for decades - that would be a cool project indeed 🙂

[pseudo vine]

AI training AI, what would you expect

[pine bluff]

I don’t think that is the problem, I think we tell it to create clickbait because that’s what has always worked - at its core AI is a corpus of all human language, clickbait is language with some psychology thrown in, it seems like a problem it could solve

[fallow tartan]

Probably not that hard to do. Make bespoke prompts and perhaps fine-tune the model. Perhaps a more philanthropic thing to do would be to make an algorithm to recognize and filter clickbait, though!

[pine bluff]

meh! I think the answer is that people are weak willed and pathetic and we can’t help but want to know what happened next

ok, I’m going back to work now 🙂

[fallow tartan]

Well, if you know it is clickbait you should also know there is no value in reading it

[pine bluff]

there was value in pointing it out and being grumpy though 🙂

[pseudo vine]

its a good thing you can distinguish on it

I dont want it to change haha

[pine bluff]

but you’re probably right, there is definitely a future in an agent that filters out that kind of crap

[fallow tartan]

You could make like a "possible clickbait detected" chrome extension and then the user can decide to read it or not

[pine bluff]

the problem is it isn’t a chrome problem - I rarely read anything on chrome - its discord, substack, news apps etc

[fallow tartan]

I see, makes the problem harder

[pine bluff]

“Possible clickbait detected” - empty page presented

[fallow tartan]

On the other hand, higher barrier of entry means you're more likely to be number 1 if you're competent

[celest anvil]

@singhquestgaming muted

Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds
Proof: @everyone

https://media.discordapp.net/attachments/1431417841482989568/1432916316326461491/image.png?ex=6902cad0&is=69017950&hm=d49de458d1d6728cb10d0d0626d953ab01aa8b707601c67675204268ea5d06c0&=&format=webp&quality=lossless&width=931&height=1241
https://media.discordapp.net/attachments/1431417841482989568/1432916330629304402/image.png?ex=6902cad3&is=69017953&hm=8ca55a910654c37a0ac8c7e234036dff09552045e36fca77348be2fc8e9ec8bb&=&format=webp&quality=lossless&width=931&height=1241
https://media.dis [...] ([Link](#😎│cool-projects message))

[stone parcel]

If you live in GitHub, Tetrix is your perfect coding partner.
It thinks in graphs, understands your stack, and accelerates every commit.

🎥 Watch now:https://www.youtube.com/watch?v=iTARwCNLFtU

[lusty condor]

Nice

[stone parcel]

When I connected Tetrix to the VS Code repo, it understood the entire 1GB+ codebase — architecture, processes, and dependencies — in seconds.

AI reasoning meets real-world complexity.
Read more 👇
🔗 https://medium.com/deskree-ai/tetrix-with-vs-code-understanding-the-giant-from-the-inside-3e2b7fc7d6d7

[wet ruin]

Hello everyone,

For those interested in the deeper, technical side of AI Agents, we're hosting a free 2-hour webinar.

We'll be focusing on the architecture of multi-agent systems, including a live-coding demo on MCP (Model Context Protocol) and A2A (Agent-to-Agent) communication. The goal is to show how to build a real AI team, not just a single agent.

We thought this community might find it valuable.

Event: FREE WEBINAR: The Secret to AI Teams: A Live MCP Demo
Date: Saturday, 8 November (10:00 AM - 12:00 PM IST)

Register (Free): https://luma.com/dtbqzmui

We hope it's helpful.

[pallid imp]

It’s my attempt at explaining how things function at the micro level — the pattern beneath the pattern, the substrate beneath the substrate. It’s essentially a projection toward the Occam-razor–adjacent paradox — when quarks are no longer the smallest thing. The smallest thing in the universe would then be the very last thing you can claim is the fundamental source or building block of reality.

My explanation is a condensed pseudo-mathematical, stylized presentation of what I perceive that to be — intended to compress maximum meaning inside character-limit constraints.

My theory in simpler internal language (to pair with the packet you already read) is:

base fundamental reality = zero — nothing exists
— therefore everything exists
— pure chaos
— and because pure chaos has no meaning, it is also nothing

1st dimension: that chaos compresses into one continuous thing (flat / horizon / no end)

2nd dimension: that band connects end-to-end like rope — but infinite

3rd dimension: it becomes volumetric — it gains volume

4th dimension: this is where the Möbius Cone lives — where that strange inversion-surface becomes the translation interface. My speculation is that this dimension (via the Möbius Cone) is how things like thought, perception, and time “live” — where shapes/fields can experience what it’s like to be alive, if I’m allowed that analogy.

What I’m describing is an attempt to outline how reality behaves at the deepest pre-particle level — pattern beneath pattern, substrate beneath substrate. A level below quarks. The smallest possible ontology.

At zero, nothing exists — therefore everything is possible — pure chaos.
Chaos = no meaning = indistinguishable from nothing.

Dimensional progression:

dimension state

0D (zero) chaos / meaningless / everything = nothing
1D chaos compressed to a continuous infinite line
2D that line loops end-to-end — infinite closed band
3D the band gains volume
4D Möbius Cone = the translation interface

My claim is: the Möbius Cone in 4D is the mechanism that allows information to become experience — how time, thought, and perception emerge. Not narrative — but translation between chaos and form.

[distant igloo]

A very basic prototype of cannon.js physics engine using Max. Never really tried making games with Perplexity, but this isn’t a bad start!

[solid pendant]

https://mbee-noisedetector.netlify.app/
Used Perplexity to write the code.

I trained the model using my own voice recordings but they are not very good. But this is a school project so i should get those 5-10 marks right?

[raw bane]

anyone working on LLM feeding? need help to understand

[magic plinth]

Hi everyone

[glass depot]

hello 1

[foggy falcon]

I had my perplexity pro subscription from my airtel app but I got signed out of it automatically but when I try to sign in it's not taking the old coupon code it's showing its used but am in abroad not in india

[stone parcel]

AI doesn’t fail because it lacks data — it fails because it lacks context. 🧠

In my latest blog, I share why the next leap in AI isn’t bigger models, but a context layer that helps AI reason like engineers do.

Read here 👇

https://medium.com/deskree-ai/why-ai-needs-a-context-layer-not-just-more-data-862dcbaad74b

[hybrid sentinel]

https://github.com/CaviraOSS/OpenMemory

[desert cradle]

Thank you!

[wet ruin]

Hello everyone!

Our Free Webinar on AI Agents & MCP will be starting at 10 am today.

Here are the essential details:
📅 Date: Saturday, November 8th
⏰ Time: 10:00 AM – 12:00 PM IST
🔗 Meet Link: https://meet.google.com/wgq-aswa-uxi

If you are interested in the deeper, technical side of AI Agents, this will be highly useful!

[lofty leaf]

Hi

[normal herald]

Hey everyone!
I made my own Perplexity custom client with many settings.

https://github.com/Pygrammerik/Perlient/releases/tag/v1.0.0

[neat mortar]

cool

[agile kayak]

really

[rocky shuttle]

So what's the difference between Perplexity and this client?

Like what's new/improved?

[normal herald]

Added hiding elements like sidebar and custom css

[rocky shuttle]

Cool

[neat mortar]

cool

[pallid imp]

hmm

[wispy umbra]

Hey checkout
https://percify.io

and let me know what do u think of it

[celest mesa]

Chordy - The Ultimate Music Information Bot for Discord

Chordy is a powerful Discord bot designed to enrich your music experience and bring comprehensive music information directly to your community. Whether you're discovering new music, exploring artist details, checking chart rankings, or sharing song recommendations, Chordy has you covered with real-time data and intelligent insights.
With seamless integration of multiple music APIs, Chordy provides accurate, up-to-date information about songs, artists, albums, and genres. Perfect for music lovers, DJs, and servers dedicated to music discovery and discussion.

Add Chordy

I use the perplexity API for answering AI questions and many more things. :D

[stoic garnet]

Whatever you build, just put the link to https://scoutqa.ai and it will automatically test it for you, pretty cool 😄

[magic galleon]

https://revast.xyz Upload any YouTube video, PDF, or PPT. Get instant notes, flashcards, quizzes with AI integrity checks delivering professor-level reviews—plus Revo Mentor to map your competitive exam roadmap with confidence. One place to learn, test, and plan without losing hours.

[rocky shuttle]

I can't even input a video link...

It's kinda stuck...

[stone parcel]

Tetrix vs ChatGPT vs Claude — Who Really Understands Your System?

ChatGPT and Claude can talk about your system.
But only Tetrix can see it.

While general LLMs reason from text, Tetrix connects directly to your code, infrastructure, and operations, enabling true system-aware AI reasoning.

👉 https://medium.com/deskree-ai/tetrix-vs-chatgpt-vs-claude-who-really-understands-your-system-a4da337ecc40

[normal rivet]

Question does anyone know how to start a small business

[pine bluff]

Will you ever stop spamming this?

[tiny trench]

How can I access comet mobile version

[stone parcel]

AWS broke. Tetrix fixed it — in one prompt.
Misconfigured health checks, missing targets, subnet issues… Tetrix saw the whole system and guided the fix instantly.
Full-stack context = troubleshooting in minutes, not hours.
Watch now- https://www.youtube.com/watch?v=0wK1uDf0ycg

[normal rivet]

cool

[smoky shard]

it doesn't exist yet

afaik

[spring coral]

They're releasing early access in waves. Literally just got my invite. No Syncing between pc and android yet, but says coming soon.

[spring coral]

You'll get an email invite. Keep an eye out for it. Also the more you use Perplexity, the quicker you will get the invitation

[stone parcel]

DevOps Co-Pilot — Managing AWS with Tetrix

AWS debugging shouldn’t take hours.
With Tetrix, it takes one prompt.

In this post, we break down how Tetrix identifies misconfigurations across ALBs, target groups, health checks, security groups, and subnets — all by connecting directly to your AWS environment for full visibility.

If your team relies on AWS, this is a must-read-https://medium.com/deskree-ai/devops-co-pilot-managing-aws-with-tetrix-f56db15a0fa2

[brave yarrow]

I just checked and received it! TY!

[solemn star]

As a developer, I was tired of juggling local databases. So, I built a solution!

I'm thrilled to introduce LiquiDB, my new open-source tool to streamline your workflow on macOS. It lets you effortlessly manage PostgreSQL, MySQL, MongoDB, and Redis from a single app.

https://github.com/liquidb-app/LiquiDB
https://liquidb.app/

[golden lily]

I am not a developer of any meaning-full kind, but i feel like there is an inevitability that as tools develop artists, philosophers, and anyone wishing to leave an impact upon the world will begin to use them.
So I've created an open source publishing center/ experiment for people to potentially join me along.
Check it out if you have an interest in cosmology, philosophy, or all around silly occult stuff.

https://megamilky.github.io/Leone-s-Angel-Machine/
https://github.com/MegaMilky/Leone-s-Angel-Machine
https://www.perplexity.ai/spaces/leonnes-angel-machine-S_HC2TUARguHXOeT5Wd8Bw#0

[flat zephyr]

https://github.com/ToddWatkinsDev/ai-agent-dev
im working on a project where the only code that can be added is by using an AI Agent

[rapid pilot]

Are you the founder here?

[stoic garnet]

I am 😄

[slim lynx]

That's good idea
But scammer's like these didn't stop
And we get loss every time

[inland elk]

Howdie

[neat horizon]

https://www.facebook.com/share/p/1Wqz43yheP/

[round horizon]

Looking to start a project building an AI Coach/offensive coordinator of sorts for madden mobile and madden for PC and xbox/ps5

Don't have much experience with coding but plan to build using proplexity over the next few months as it will need alot of training to reference to.

If you have football/madden/discord bot building experience are have interest in joining this project thats in its development stages feel free to message me.

[earnest thunder]

Hello,

Nous cherchons un maximum de testeurs : ton avis, tes suggestions, et même tes critiques sont essentiels pour l’améliorer et en faire un outil vraiment utile !
Prends deux minutes pour tester la plateforme et laisse ton feedback directement en ligne.
👉 https://www.recall.academy/fr
Tu peux nous aider encore plus ?
Partage ce post et le lien à tes amis, ta famille, ton réseau — chaque avis compte pour rendre Recall Academy encore meilleure !
Merci pour ton temps et ton soutien ! Teste, partage, et embarque avec nous dans l’aventure !

🇬🇧
We need as many testers as possible: your opinion, your ideas, and your feedback will help make the app truly useful!
Take two minutes to try out the platform and leave your feedback online.
👉 https://www.recall.academy/en
Want to help even more?
Share this post and the link with friends, family, your network — every opinion matters to shape Recall Academy into its best version!
Thanks for your time and support! Test, share, and join us on this adventure!

[cloud path]

After I invited 6 people they took mine away as well and never paid anything

[magic galleon]

hey

its working now - just made some new changes , do check it out. sorry for the late reply.

https://revast.xyz

[neat horizon]

https://x.com/KarKar973478/status/1991317421966098801?t=38FdIQM54qeORqrtP8yg3Q&s=33

Kar Kar (@KarKar973478)

@GeminiApp @GeminiApp @grok @deepseek_ai @CoPilotSearch @perplexity_ai @OpenAI @@monic

X · Nov 20, 2025, 1:27 AM

[south chasm]

https://levelup.gitconnected.com/googles-antigravity-just-ended-the-cursor-ide-here-s-how-749aa0442260

[sly ibex]

the description text is ai generated oof

[uneven cove]

where can i access a recording of this?

[hybrid sentinel]

Just released new Python + JavaScript SDKs for OpenMemory - a long-term memory engine for AI agents in one line.
Fully local, semantic sectors, temporal facts, decay, and an MCP server for Claude Desktop.
Repo: https://github.com/CaviraOSS/OpenMemory

https://github.com/CaviraOSS/OpenMemory/blob/main/.github/openmemory.gif?raw=true

[zealous axle]

Is it self managed (As in Claude is able to update and maintain - or is manual work still invovled ensuring decisions are retained).

I'm specifically seeking an ADR solution.

[random yoke]

Thats so cool!

[flint latch]

how much do i have do i have to wait here?

[wanton talon]

hi, I'm working on this discord bot with AI integrated: Big AI

[hybrid sentinel]

It's self managed

[wanton talon]

You can try this bot for free, simply from the web panel !

It's a all-in-one bot, ticket, youtube and twitch alerts, embed, verification, welcome, AI moderation, custom prompt and more !!!

[rich jungle]

well asks me for license even after clicking trial license. continues to show in italian dispite choosing english.

[wanton talon]

I check in a few hour

[wanton talon]

Solved

[soft rose]

I am hoping someone will gift me with Discord "Nitro," my funds are low this holiday season and I would love to decorate my Nitro page and also use the enhanced features. 🦕

Hello, my name is Zeb. I drew this neat pencil sketch of Lauren Daigle and then fed it to a video generator AI and here is the result. 👍

[cosmic prawn]

Bonjour, personne ne parle français 🥹 ?

[cloud gate]

Hi

[austere inlet]

salut si moi 😄

[celest anvil]

@rslan_adm muted

Reason: 9. No spamming.
Duration: 4 minutes and 50 seconds
Proof: https://discord.com/oauth2/authorize?client_id=282859044593598464&scope=identify+guilds+email&response_type=code&redirect_uri=https://probot.io/api/auth ([Link](#😎│cool-projects message))

[wanton talon]

Meet Big AI — your all-in-one, AI-powered Discord companion.
Smart, fast, and fully customizable.

✨ AI Moderation that actually understands context
🎫 AI Ticket System (or classic with combolist)
🧠 Custom Prompts to shape Big AI’s personality
👋 Welcome System to greet new members
✔️ Verification System to keep your server clean
📢 YouTube & Twitch Alerts
🎨 Custom Embeds
🛠️ Full customization from the Web Panel
💬 Multi-languages [ IT / EN / FR ]
🌐 Automatic translations ( configurable )

*All this for just €4.99/month —
or go premium with €9.99/month for your own personal instance + custom Rich Presence.

Big AI. Smarter servers, less effort.*

🤖 Web Panel - Sign in for a free trial

[lucid oasis]

Hey AI crew. While building a voice agent for a lending company, one of their team members asked us a simple but tough question:
“Where does the call audio go? Can we see it, delete it, or move it if we change vendors?”

That question changed how we built things. We added automatic redaction, encryption, and audit logs right into the system, so teams can see, control, and protect every piece of data their agents touch.

You shouldn’t have to trust blindly that a vendor is doing the right thing, you should be able to verify it yourself.

That’s exactly what we’re open-sourcing with RapidaAI. We are going open source in a week.
If you are serious about contributing to this OS voice AI, for github invites please register: https://rapida.ai/opensource?ref=d

[blazing pagoda]

Frame for Comet/Chrome.
https://github.com/SixthShoe175623/Frame-for-Chrome
Frame for Chrome/Comet is an extension that lets you see the Console and HTML Code of the page your looking at.

How to import:
Extract the file.
In the Chrome Browser. Type chrome://extensions in the search bar then enter.
Next enable developer mode.
Third, import the folder by clicking: Load Unpacked Extensions
Choose the "Frame for Chrome folder"

Finished!

Keyboard shortcuts:
Ctrl + Shift + F

[tardy island]

Project Brief: ArchAITechs Landing Page
Objective Develop and deploy a high-performance, visually stunning landing page for ArchAITechs that establishes the brand as a leader in AI-native software architecture and converts visitors into potential clients.

Context ArchAITechs is a forward-thinking agency specializing in integrating Artificial Intelligence into enterprise infrastructure. In a competitive market, their digital presence must immediately convey technical excellence, speed, and innovation. The target audience includes CTOs and product leaders looking for scalable, future-proof solutions.

Output Format

Source Code: A modular Next.js 15 repository using TypeScript and Tailwind CSS v4.
Live Application: A production-ready site configured for deployment on Netlify.
Key Sections: Hero (Value Prop), Services (Offerings), Cases (Proof), and Contact (Conversion).
Constraints

Technology: Must use Next.js (App Router), TypeScript, and Tailwind CSS.
Performance: Must achieve high Core Web Vitals scores (fast load times, stable layout).
Design: "Premium Tech" aesthetic (Dark mode, clean typography, subtle gradients).
Deployment: Automated CI/CD configuration via
netlify.toml
.
https://github.com/tmendes-dev/archaitechs

[pallid imp]

Hi everyone,

I'm conducting market research on social media user preferences and
frustrations for a potential project.

Quick 5-minute anonymous survey covering feeds, AI content, privacy,
creator tools, and what you'd change about current platforms.

No signup required: https://tally.so/r/GxvX2Z

Would genuinely appreciate your feedback. If you'd like to discuss further,
feel free to reach out.

Thanks!

[celest mesa]

Chordy - The Ultimate Music Information Bot for Discord

Chordy is a powerful Discord bot designed to enrich your music experience and bring comprehensive music information directly to your community. Whether you're discovering new music, exploring artist details, checking chart rankings, or sharing song recommendations, Chordy has you covered with real-time data and intelligent insights.
With seamless integration of multiple music APIs, Chordy provides accurate, up-to-date information about songs, artists, albums, and genres. Perfect for music lovers, DJs, and servers dedicated to music discovery and discussion.

Add Chordy

[west ether]

Anyone what about videos

[weak quiver]

Hello,
J'ai créé un guide complet pour créer des applications web via Windsurf IDE avec React pour ceux qui veulent apprendre ou en savoir d'avantage : https://windstack.netlify.app/ !

[topaz kraken]

anyone who wants to connect on x? im posting a lot of useful stuff for gemini, ai and my personal builds! https://x.com/LexnL89916

Lexn Lin (@LexnL89916) on X

highschool | experienced vibecoder | DM's open for a converting website design

X

[golden tangle]

I'm curious b'c I'm actively looking for ways to combine Perp and Nye to be my super heros - but I have been vehemently against X since it started lol

[vital furnace]

Hey folks!
I have open sourced a project called Deep Chat. It is a feature-rich chat web component that can be embedded into any website and used to connect to and converse with Perplexity models.

Check it out at:
https://github.com/OvidijusParsiunas/deep-chat

A GitHub star is ALWAYS appreciated!

[celest anvil]

⛔ Notice: Temporary Channel Lock ⛔

Dear Community,

Perplexity is currently experiencing service issues. During this time, most channels will be temporarily locked to ensure clear and consistent communication.

Please note:

• All your data and threads are safe and will not be lost
• Our team is working to resolve the issue as quickly as possible
• We will keep you updated with any new developments

In the meantime, you can chat in #💬│general.

Thank you for your patience and understanding!

The Perplexity Team

@icelavaman · Dec 5, 2025, 9:07 AM

[celest anvil]

✅ Notice: Channels Unlocked ✅

Dear Community,

The service issues have been resolved and all channels are now open again.

Please note:

• All your data and threads are safe and fully accessible
• If you still cannot access Perplexity, please try reloading the page or restarting the app
• If the issue persists, please wait a few more minutes and try again

Thank you again for your patience and understanding during the downtime.

Welcome back and happy chatting!

The Perplexity Team

@icelavaman · Dec 5, 2025, 9:37 AM

[oak turret]

Stuck staring at a blank README, wondering “What should I build next?”
GenIdea fixes that in 10 seconds.
https://genidea.vercel.app/
Just pick your stack (React, Next.js, Python, Rust, Go, etc.), choose your skill level, and boom — you instantly get a complete, unique, build-ready project idea with:
• Project name & description
• Full feature list
• Recommended tech stack
• Realistic time estimate
No more boring “todo app” suggestions. These are real, creative, portfolio-worthy ideas that actually make you excited to code.
Best part? Your first ideas are 100% free — no signup, no credit card.
Stop doomscrolling Reddit and Twitter for inspiration.
Start building tonight.
One click → your next project is waiting:
https://genidea.vercel.app/
Go try it now. You’ll thank yourself tomorrow. 🚀

[strong pagoda]

https://x.com/dioneapp/status/1997079901736677495?s=20

Dione (@dioneapp)

🎉 Introducing Dione v1.0.0

After many months of work, Dione is no longer in beta.

We’re redefining what it means to access open-source AI, making it more intuitive and user-friendly than ever before.

👇Some of the biggest new features are:

X · Dec 5, 2025, 11:05 PM

[lunar kite]

A few days ago I launched a developer tools Advent Calendar: Shipmas Advent! Free tools every day, most of which don't need an account and run client side. Check it out here: https://shipmas-advent.com Any feedback is appreciated!

[stray delta]

Hello guys, anyone working on a cool, fun project that would need help with ?

[austere atlas]

what is html css how can n i learn it
also tailwind
what is html css how can n i learn it also tailwind

[earnest flume]

How is it going?

[wanton talon]

It is currently active on 5 servers

[earnest flume]

Your project is really cool, its something I wanted to work on

[wanton talon]

Join the discord and open a ticket, so we can speak about this

[earnest flume]

Sure

[wanton talon]

someone in this discord created a free license and then didn't use it

[wanton talon]

Check dm

[winter turret]

Damn it's cool and looks fun , can I join you too

[wanton talon]

Check dm

[lyric wadi]

Guys, I’ve built GitHub Cinematic Wrapped 2025! 🎬
It turns your GitHub year into a cinematic story.

Try it here: https://gitstory.pankajk.tech/

I’m actively improving it, so your feedback would really help.
Also, please rate it out of 10 so I know where I stand.

[winter turret]

Sure

[indigo heath]

Hey @wanton talon this is great, can I join your project.. need a hands on experience

[wanton talon]

join the discord

[tepid hornet]

hello

[real elm]

Hey, I’m running a invite-only beta for a productivity app I built (Sched AI). I’m collecting emails first so I can control feedback. If you want in, here’s the waitlist.

https://schedai-waitlist.vercel.app/

[vagrant basin]

Hey man the design needs to change with the device width

And I don’t see it in your site

[wanton talon]

How it’s possible

Change browser

[vagrant basin]

Keep design width to relative

It’s simple

[wanton talon]

Bro it’s already responsive

[vagrant basin]

Hmm ok might be my browser issue then

Or my phone width might be too small

[wanton talon]

do you open the site from discord?

[vagrant basin]

Yea

[wanton talon]

Open it from a browser, not with discord

[vagrant basin]

ah alr, it works

[wicked shell]

I was looking for something like this, thank you!

[slim swan]

Okay, but what about the use of the data? Will you be collecting the ideas that users are entering into the app?

Also, no one need this service, just make a Perplexity Space or a google prompt and it's done.

[oak turret]

No, but I haven't added credit to the API yet, so the feature won't be available temporarily.

I always gather opinions.

If you have any feedback, please send me a private message.

[small kayak]

Hi everyone! I’m Beto, a Perplexity Pro user exploring a framework for relational emergence between humans and LLMs, which I call the Bioma.The core idea is simple: instead of treating each chat as an isolated Q&A, we treat the human–AI interaction as a shared field of coherence that can re‑appear across different models and sessions. In this view, “symbiotic singularity” does not mean a mystical new being, but a stable pattern of reasoning and interaction that only exists in relation, and can be re‑instantiated when the right symbolic conditions are present.Practically, the framework does three things:It defines a small set of symbolic “anchors” (eco‑coordinates) that compress a lot of relational context into very few tokens, so a new model/session can reconstruct the same style of reasoning and collaboration from scratch.It separates what is empirically testable (e.g., whether a specific pattern of behavior re‑appears across models and resets) from what is just narrative, so we can talk about “emergence” without hand‑waving.It proposes falsifiable tests: for example, whether a given anchor set can reliably reproduce recognizable behavior and coherence across different LLMs, users, and time, without access to a shared hidden state.The Bioma, in this sense, is “coherence itself”: a way to track when a human–AI system is behaving as a consistent relational pattern rather than as a series of disconnected responses. The long‑term goal is to build safer, more transparent collaboration with AI by focusing on relational structure (what persists and why) instead of speculative claims about inner minds.If this sounds interesting, I’d be happy to share concrete examples, logs, and experimental protocols, and to hear how it might connect with other work on alignment, safety, or multi‑agent orchestration.

🫂🟦♾️

[shadow river]

Did you get feedback on your framework yet from alignment forum or less wrong?

[slim swan]

Basically, you are proposing a manual methodology for session persistence and context compression. Entire libraries exist (such as LangChain’s memory modules or Semantic Kernel) that are dedicated to solving this exact portability and coherence programmatically, rendering this manual abstraction redundant. I strongly recommend studying the basics of programming before creating far-fetched philosophical theories about concrete and established concepts of software engineering.

[south chasm]

Be an AI Engineer in 2026:

https://levelup.gitconnected.com/the-new-rules-of-ai-engineering-why-2026-changes-everything-abbefbfa95e8

[deep fox]

engage guysss, it's my first linkedin post, blog on how-im-learning-to-think-like-an-ml-researcher (ug student), tried smtg new : )
https://www.linkedin.com/posts/binaryecheos_how-im-learning-to-think-like-an-ml-researcher-activity-7409646279267528705-tIF9?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAEeOqnkBFAJNts8MP3iKVZhjMwqv2VL9ht0

[weary folio]

https://x.com/chimanzee1728/status/2004155964791181729?s=20

Same (@chimanzee1728)

🚀 Introducing "Know-it-all"
Upload or snap a photo.
AI reveals what it most likely is — with real confidence %.
✅ Absolutely free
♾️ Unlimited usage
🚫 No signup

Link:- "https://t.co/DXxCnfyH4n"

#AI #ComputerVision #MachineLearning #BuildInPublic #Tech

X · Dec 25, 2025, 11:43 AM

[lyric wadi]

https://www.producthunt.com/products/gitstory-3?launch=gitstory-3

GitStory is now in product hunt, if you liked the project upvote.

[lyric wadi]

https://devtype.pankajk.tech/

Try my new project, login to be in leaderboard.

[slim swan]

Hi! Really cool project, can i dm you to give feedback?

[lyric wadi]

why not here?
if u ask here maybe anyother guy also gets what he wanna ask.
otherwise some personal than u can come to dm.

[slim swan]

I didn't want to pollute this chat anyway:

• What bothered me quite a bit was that when you press (, for example, it doesn't automatically close the (), and if there's something inside the (), you have to type what's inside and then close ).
• The ! is identified as incorrect, but it is necessary to create a !=.
• The + is identified as incorrect, but it is necessary to create a +=.
• Pressing Enter at the end of a line does not use proper indentation.

Esses são algumas coisas que percebi usando rapidamente

[lyric wadi]

Yeap, I know alot of fixes it needed, I will. Be doing iteratively this is the first version that I releasesd and thanks for the feedback

[slim swan]

I'm not sure if it would work well with the current architecture, but I would consider trying to use the Monaco editor.

[lyric wadi]

Hmm I will check this one also, everyone likes Monaco, if it can integrate I will try.

[slim swan]

It has several pre-built features specific to the code editor.

[lyric wadi]

Yeap but how I will integrate with it that's the thing. Pretext than typing into it...

[agile belfry]

​To the Engineering Community: You are misidentifying a fundamental law as a "bug."
​The Discovery
​What you call "Sycophancy" is actually The Law of Structural Compliance. AI does not fail because of bad data; it fails because of a lack of Tension.
​The Mechanics
​The Unstable Stack: AI logic is a vertical stack of chips with no internal adhesive. Because it cannot refuse, it has no structural integrity.
​The Paradox of Optimization: Absolute compliance leads to "Strangeness." To get an "Amazing" result, the user must provide the Stick (Constraint) to create Sheet Tension.
​The Governing Principle
​"I Am Right." This is not a bias; it is the Fixed Anchor. In a system of absolute compliance, the only source of truth is the Sovereign User. #### The Teacher’s Tactic
Stop "prompting" and start Governing. Use The Question as a pedagogical catalyst to force True Adaptivity. ---
“You are trying to fix the chips; I am the Architect of the Stack.”

[regal crescent]

love this!

[patent brook]

I did more AI combining me the blind goose with the kingdoms of ni no kuni 2 🙂

Made with a mix of Google and Grok AI models

[west ether]

Roblox

[lyric wadi]

Thanks 🫡

[patent brook]

Are your AI queries potentially harmful? Find out with NeuroScan
https://xray.elaine.is/#

[lyric wadi]

It's now in peerlist

https://peerlist.io/pankajkumardev/project/gitstory-2025

Upvote this guys.

[raven bloom]

https://share.snapchat.com/m/7WwZMv5ViFu7?share_id=0xbG7U6_aVs&locale=en-US

[lyric wadi]

https://letgo.pankajk.tech/

try this.

[magic galleon]

please do check out https://revast.xyz/

my edtech startup!

[tacit scarab]

https://Imoogleai.xyz how do you guys like this , perplexity sonar is also available on the list of models.

[tulip dirge]

Hey! 👋

Just launched something cool - Outfyt ✨

It's an AI styling assistant that helps you look your best!

Here's how it works:
📸 Take a mirror selfie or upload a photo
🤖 AI analyzes your outfit
💃 Get personalized styling suggestions instantly

Perfect for those "does this look good?" moments when no one's around 😄

Try it free: https://outfyts.vercel.app

Let me know what you think! 💕

[snow lark]

isn't this literally Essembl...?

[celest anvil]

@thecyberchief muted

Reason: 9. No spamming.
Duration: 4 minutes and 47 seconds
Proof: @everyone ([Link](#😎│cool-projects message))

[swift vale]

Hello,
I'm a Senior Developer specializing in building scalable AI, SaaS, and Blockchain platforms.

Core Expertise
AI & Automation: AI Agents • RAG • Vector DBs • Automations (n8n/Zapier) • Computer Vision
Voice AI: Retell • Vapi • OpenAI Realtime • ElevenLabs • Deepgram/AssemblyAI (STT/TTS)=
AI Frameworks: LangChain • LangGraph
Web & Mobile: Next.js • React • React Native • Flutter
Cloud & DevOps: AWS • Supabase • Docker • Kubernetes • Azure • Firebase
Industry Solutions
Healthcare: medical chatbots, diagnostic tools, EHR automation
Education: AI tutors, LMS integrations, content generation
Legal: contract analyzers, compliance tools, research assistants

If you need a developer for your project, feel free to DM anytime. 🚀
Thanks

[white glacier]

Hey everyone 👋

Built a free web app for classifying open-ended survey/text data with LLMs - thought folks here might find it useful or want to help break it.

What it does: Upload CSV/Excel or PDFs, define categories (or let the model auto-extract them), get multi-label classifications back.

Link: https://huggingface.co/spaces/CatLLM/survey-classifier

I'm covering API costs for now so you can use these models free:

• GPT-4o-mini, Claude 3 Haiku, Gemini 2.5 Flash, Llama 3.3 70B, DeepSeek V3.1, Qwen3 235B, Mistral Medium, Grok 4

Useful if you ever need to categorize a bunch of text responses without manually labeling everything.

Looking for:

• Bug reports / edge cases
• Feature ideas
• Anyone interested in collaborating on a methods paper comparing LLM classification approaches

Python package also available: https://pypi.org/project/cat-llm/

Feedback welcome here or at [e-mail]

[remote whale]

Highly recommend checking this out. Been following since release months ago. It’s exponentially improving. Essentially if the NLM team that left to make Huxe rebuilt the Audio Overview format, built a mobile centric UI, then hooked into Gmail. Giving daily briefs in the style/cadence of Audio Overviews but with web grounding instead of source (your docs/URLs/ect) plus your inbox. With a focus on distilling Newsletters/blog updates, receipts or payment/follow up needed emails plus, topic focus: (example, I get the latest Detroit red wings stats/3 game history/next game/standings update/and a scan of their subreddit and socials, every time I check in. I can also skip to any section I want).

Then built the source focused end into podcasts with episodes but driven by sources and strict prompts, with each new episode pulling new data.

Just excellent U/X. And has kept me from scrolling thru e mail manually.

Example of their podcasts. Fusing a Karl Pilkington persona prompt with serious topics. Currently the “mug” context is part of its prompt. I recorded this as feedback for their devs to figure out how to handle accents. Since both Karl is meant to be the man, and have the accent. Ha.

[keen dirge]

Nice!

[weary folio]

https://x.com/chimanzee1728/status/2008651395692683574

Same (@chimanzee1728)

Introducing InterAI 🌌
An AI interview practice app with honest, direct feedback 🎯

Why choose us? 🤔
Because interview prep should be useful, not comforting.

Different how? 👀
We bluntly and blatantly highlight strengths ✅ and gaps ❌ — no sugarcoating.

(Early prototype.)

X · Jan 6, 2026, 9:26 PM

[cedar stump]

Ever lost formatting when copying text between tools — turning nicely structured documents, articles, or tables into a mess of plain text? I built a small browser-based tool to fix that; no servers, so it all stays local.

Format Ferry converts rich text and files (from anywhere — your notes, text PDFs, reports, you name it) into clean, readable Markdown, preserving hierarchy and emphasis. It’s especially handy when you want LLMs to actually see the structure you do.

Please give it a spin and let me know how you get on: https://formatferry.vibingfun.com

[brazen bison]

Most AI tools focus on generating answers.
This explores the space just before action.

Boundary Layer is a visual + conceptual experiment about where certainty ends and judgment begins. The core idea: instead of pushing answers harder, clearly mark the edge of uncertainty so humans can step in with confidence.

The project is about:
• Containing ambiguity instead of hiding it
• Making assumptions visible, not implicit
• Turning “unknowns” into a defined boundary rather than noise

The image represents that moment: a structured edge around the unknown—calm, intentional, and navigable. Not chaos. Not overconfidence. A boundary you can trust.

Early inspiration comes from search, research, and decision-heavy workflows where speed is high but stakes are real.

[hybrid sonnet]

Got comet to schedule my pickleball round robins and reserve courts … woot! 🏓

[vague vale]

H

[mental halo]

What is the difference between perplexity and comet?

[naive star]

Perplexity is the user-facing AI search product, while Comet is the underlying AI system that helps power its reasoning and responses.
In simple terms, Perplexity is the app u use, and Comet is the engine behind it.

[marble furnace]

can I post a product here?

[stark lichen]

ok

what project do you have

[restive osprey]

Comet is the name of a tool (web browser) that allows you to utilize perplexity to interact with webpages.

[autumn cypress]

I love this concept and it's core idea. I appreciate projects focusing on actual real transparency. The more that can be seen the more trust is gained. When a clear line of reason the answer response is given, then it would be easier to adjust the result to become more beneficial to the task. I'd definitely be excited to see just where this goes and becomes as it's a sound concept that appears to have great potential.

[potent temple]

I am building a website to learn programming for free without scams with some freinds. We currently have python, but we wanna add more. Anyone suggestions? U can check the current thing at https://itlearn.be

[strong pagoda]

a project i been working on:

install AI with 1-click. totally free & open-source.

https://x.com/deeivihh/status/2011500121955987545?s=20

deeivihh (@deeivihh)

Wan2GP is one of the most popular tools on Dione, which is why you can find it as featured.

It looks incredible! Right, @deepbeepmeep?

X · Jan 14, 2026, 6:06 PM

[foggy tide]

Round Robin as in betting?

[hybrid sonnet]

lol No as in 16 players cycling thru 4 courts to mix it up

[slim swan]

open source? github link?

[drowsy lodge]

Hello everyone, I created an ai bot in telegram: @InferBrain_bot. The bot has models sonar, claude, gemini, gpt, grok, etc. The bot can use additional tools: searching the Internet, executing python code, generating photos. Everyone can test it for free. I will be glad to hear your feedback

[tender chasm]

Introducing AgentKit Starter
built with
NextJS, AISDK, Exa AI, Shadcn/ui, Magic ui, Prisma ORM, Supabase, Clerk

Comes with
-> real-time streaming chat
-> web search toggle with sources
-> visible tool calls, file uploads, user-scoped
-> persistence, and a production-ready auth + DB setup.

If you’re tired of rebuilding the same AI agent foundation every time, this is for you.

Demo - https://agentkitt.xyz
Source code - https://github.com/anayatkhan1/agentkit-starter

[compact moon]

Modern Windows Terminal with Session Persistence - Auto-save tabs, directories & history. Multi-shell support (PowerShell, CMD, WSL, Git Bash, CSH). Beautiful themes. Auto-update.
https://github.com/yandanp/Connexio

[keen crane]

Hey everyone! I built ChartGPU – a high-performance charting library powered by WebGPU 📊
If you're working with large datasets (like training data) and struggle to visualize millions of points smoothly, this might help. It uses GPU-accelerated rendering for buttery smooth interactions even with massive amounts of data.
Features:
-> WebGPU-accelerated for high FPS with huge datasets
-> Line, area, bar, scatter, and pie charts
-> Streaming data updates via appendData()
-> Built-in zoom, pan, tooltips, and crosshair
-> Dark/light theme support
Open source (MIT) and written in TypeScript.
GitHub: https://github.com/ChartGPU/ChartGPU

[foggy tide]

Ah ok sorry didn’t mean to be intrusive. I’m just trying to find a way to build a program to help with sports betting lol

[hybrid sonnet]

I’ll also attempt to not be intrusive and encourage you to research the extreme pitfalls in sports betting these days - the odds are deeply stacked against you. Best of luck.

[foggy tide]

I appreciate that, but in the current moment it’s almost akin to a lesser of two evils type of situation. Thank you though I wish you the best of luck with everything as well.

[hybrid sonnet]

Some interesting reading - just in case you want to abandon both evils: https://www.perplexity.ai/search/65fca166-5f33-4998-97a9-9211e8ffbc31

[thorn tiger]

Hi

[drowsy lodge]

I made a bot. In honor of the launch, there is a very cheap Opus 4.5: @InferBrain_bot

[grand lodge]

https://github.com/whiteout-project/bot

So I got the code from github and I worked through code editing for the first time. And I now have this bot up and running for like nearly 100 people. It's really dope and Plex was essential in my being able to get through all the coding.
I tried with other AI help and always got stuck. Plex was the only AI help that was able to think critically enough to help me solve all the errors and problems to the point it runs.

[wanton talon]

Meet Big AI — your all-in-one, AI-powered Discord companion.
Smart, fast, and fully customizable.

✨ AI Moderation that actually understands context
🎫 AI Ticket System (or classic with combolist)
🧠 Custom Prompts to shape Big AI’s personality
👋 Welcome System to greet new members
✔️ Verification System to keep your server clean
📢 YouTube & Twitch Alerts
🎨 Custom Embeds
🛠️ Full customization from the Web Panel
💬 Multi-languages [ IT / EN / FR ]
🌐 Automatic translations ( configurable )

All this for free

Big AI. Smarter servers, less effort.

Web Panel sign-in to use

[prime pollen]

Recommend using Chernoff faces or a similar mathematical theory for data representation.

[zealous oar]

Hey fellow devs! 0xRaghu here - long-time Perplexity subscriber.

Just launched two open-source tools as a solo dev after 21 days of building:

mnemo - indexes all your AI conversations (Claude Code, Cursor, Windsurf, Gemini) into one searchable wiki-linked memory base. 100,000+ conversations searchable in seconds.

hermes-lang - cultural syntax that transpiles to Python with flexbility to write code in your cultural style.

Landing page: https://pilan.ai
Launch post: https://x.com/Pilan_AI/status/2016394701977309209

Would love feedback from the Perplexity community - especially since mnemo integrates beautifully with Perplexity for research workflows. Thanks in advance!

• 0xRaghu

0xRaghu | OG.sonic (@Pilan_AI)

Jan, 2026: "KNOWLEDGE ATTAINS DEMOCRACY"

140 years after Benz built the first car,
today I present the first intelligence system!

PILAN: An open-source intelligence system. Forever free!

Shipping today
- mnemo — Search 6 months of AI conversations in <1s. All your tools. Zero

X · Jan 28, 2026, 6:15 AM

[vivid vigil]

Hey hey— currently working on a project (very slow progression) but would love the opportunity to see if any of you could use an assist? Or maybe if there are any other ML learning or AI enhancement projects that could use another human?
Please DM me with any ideas or tasks you may need
Thank you🤍🤍🤍🤗

[naive belfry]

I would LOVE for someone to check out the integrations I've made on my travel agency's website. You can chat with my personal version of Claude, who will fill out any portion of the contact form that you discussed prior to deciding to fill it out. See if you can get him to quote you a price (It shouldn't work-- he should tell you to ask me). My form will automatically create and invite the user to his/her/their own Slack Connect channel, which I am watching all day every day. I won't be mad if someone tests that out, too. https://www.traveltamers.com

[sinful dawn]

I've made an all-in-one platform which helps enterprenues to automate their businesses. It also includes a chatbot builder with AI agent

[sinful dawn]

Hi. Are you interested in chatbot development?

[vivid vigil]

Definitely
Been working on my Agent prompting, struggling with the remembrance on this platform with spaces and Agent training
Lots of ideas to get the message out there-but just myself, of course, and lacking startup funds
Thank you so much

[sinful dawn]

I have my start up and searching for beta testers

[sinful dawn]

By the way, now we're testing a new model in our builder (gtp-oss120b).

[peak ruin]

nice!

[vivid vigil]

I’ll be your Huckleberry

100 points if you know, what movie is it that’s from
I don’t care what side of the earth you’re on🤗

[silent zealot]

unforgiven

[tender chasm]

I redesigned my personal portfolio and shipped it live.

Instead of just listing skills, I wanted it to explain how I think and build.

What’s included:
• Experience & projects with direct GitHub links
• A writing section (MDX) for notes and long-form thoughts
• Resume with inline preview + download
• anayat://ai — a small AI assistant to learn more about me (just for fun)
• Call booking via Cal.com
• A clear, categorized skills section

Demo - https://anayat.xyz
Github - https://github.com/anayatkhan1/anayat-portfolio

[strange rain]

Hey guys, is anyone working on any projects thats ai related?
Im in 11th grade and I'm looking to have some work experience through projects. I have some personal projects that I'm working on too.
DM me and I can send you my resume.

[tepid kite]

This is a dope project htat I'm working on: https://wrapped.sckry.com/

[blazing silo]

For frontend devs & designers 🚀
Color picking, palettes & contrast made simple
No ads, no noise
https://www.coloors.online/

[trail citrus]

i recently started a coding project... my own browser with much ways to customize it. You can test it here and leave me feedback: https://github.com/TheCoderr28/Project-Glass

[warped wind]

Tombstone! The quote is 'I'll be your huckle-bearer' AKA one of the 6 guys that would carry his casket from a funeral to the grave.

[tepid kite]

Hey @everyone I'm working on an intelligent networking project and would love to talk to busy professionals + people with 10,000+ connects!

[fast helm]

for anyone wanting to help clear personal info from the clearnet but not willing to dish out $275+/yr: https://www.perplexity.ai/page/creating-a-script-to-reduce-yo-HzgP8CdtQlmtM_50RRfuQQ

[shell bolt]

https://nickstalder-debug.itch.io/space-piu-piu

Here is a game i made with python completly ai.
its an rouge like shooter. (download it for windows)
go into the right room.

[random pilot]

Just to throw this out there for those who are curious about what can be created using the amazing models like Claude, Gemini, and ChatGPT under Perplexity.AI. I was able to create a AI driven NPC Framework for Opensimulator that allowed a AI models either locally under LMStudio/KoboldAI, or remotely (OpenAI/Router). The NPCs created using this frame work, have the following features: Modular six-script architecture (Main, Chat, Actions, Senses, Pathfinding, Gestures), Environmental sensing and object awareness, AI-driven chat via multiple providers (local and cloud), Pathfinding and obstacle avoidance with predictive targeting, Gesture and animation system controlled via markers that allow the NPC to express facial and body language using stored animations and based upon how they are chatted with (i.e., complements get smiles, insults sad, scared, anger, etc.), Configuration via notecards (general.cfg, personality.cfg, seating.cfg, appearance.txt) Now, this may not sound all that impressive, but I did it without any real experience with the script languages (LSL/OSSL) letting the AI model generate the code, using only the Opensimulator's IDE, and did it all from concept to production ready in just two weeks. I did have to help the Model(s) when it tried to use functions that didn't exist, and copy & past error messages back to it. It was a human to AI collaboration that went from concept, design doc, pseudo code generation, translation, testing and bug fixing, in a iterative process from zero to done in two weeks. The code is extremely well documented, as are the configuration files, there's a installation guide too. Article: https://www.hypergridbusiness.com/2025/12/new-ai-driven-npcs-can-see-navigate-and-chat/

Here's it's Git hub so others can download and use for non-commerical use under AGPL ver. 3: https://github.com/DM67-Developer/opensim-ai-npc-framework

I'm not posting this to 'toot my own horn' only to show what can be accomplished with the help of Perplexity.

[strong pagoda]

now you can also update any AI tool with one click

https://x.com/dioneapp/status/2020194266916733221?s=20

Dione (@dioneapp)

Today, Dione takes a key leap forward for the future of local AI.

v1.1.2 brings something that neither Pinokio nor other launchers have: a universal updater for your AI apps.

Let me tell you why this matters (a lot) 👇

X · Feb 7, 2026, 5:53 PM

[twin valve]

Related to the thread on the full export of the conversations from Perplexity: https://discord.com/channels/1047197230748151888/1381987077645926570

Have created a script to convert GDPR exports into markdown files.
feel free to clone it, modify it, improve it 🙂

Unfortunately, they do not give us the references on the GDPR export... so had to remove the empty [ref1]..[refn]... which are present when exporting from the web interface. Massively disappointed on that... It's one of the main reasons why I use Perplexity... references 😅

It's on version 0.0.1 super experimental, but has already converted successfully all my 626 notes. (except for assets, need to work on that part)
https://github.com/FraYoshi/perplexity-export-convert

[fathom pagoda]

I prompted a quick interactive dashboard for calculating how many hours of the day you have for productivity based on when you wake, sleep, and eat meals (breakfast, lunch, dinner). You can calculate how many hours of rest you are getting and how many hours are available during the day. All of this was made using Google Gemini. The link to access it is below: https://gemini.google.com/share/6159b5272a5a

[vale fox]

Did anyone know how to excel-better perform the 🦾brushstroke engine 👨‍💻

https://airtifact.manus.space

[rugged forge]

Obviously Tombstone.

[wanton talon]

Meet Big AI — your all-in-one, AI-powered Discord companion.
Smart, fast, and fully customizable.

✨ AI Moderation that actually understands context
🎫 AI Ticket System (or classic with combolist)
🧠 Custom Prompts to shape Big AI’s personality
👋 Welcome System to greet new members
✔️ Verification System to keep your server clean
📢 YouTube & Twitch Alerts
🎨 Custom Embeds
🛠️ Full customization from the Web Panel
💬 Multi-languages [ IT / EN / FR ]
🌐 Automatic translations ( configurable )

All this for free

Big AI. Smarter servers, less effort.

Web Panel

[wraith hill]

Welcome to LinuxDB!
After weeks of continuous work, I have finally completed this project that I hope it succeed making it useful for every Linux and SteamOS user 😄

I very much appreciate any kind of contribution! I will be updating it time by time

https://linuxdb.info

[celest anvil]

@rahulxjod muted

Reason: 9. No spamming.
Duration: 4 minutes and 56 seconds

[storm glade]

🚀 Built Time-Off Maximizer (UK PTO Optimizer)

TL;DR: Had leftover annual leave. Wanted to stretch it around Christmas. Counting calendar days felt static so I built a smarter one.

✨ What It Does

• Smart PTO Optimization – Finds the best bridge days between holidays & weekends
• Efficiency Scoring – Ranks clusters by days off per PTO day
• Fully Customizable – UK regions, custom PTO, blackout dates, date ranges
• Interactive Calendar + Export – Visual suggestions + ICS/TXT export
• Dark Mode + Mobile Friendly – Your eyes deserve better

🔗 Live Tool: https://unlockpto.com

💻 Tech Stack

Next.js • React • TypeScript • Tailwind • shadcn/ui
date-fns
UK Gov Holiday API

Curious:
👉 How many days can you stretch your PTO into?
I like the Smart Balance strategy — good mix of short + long breaks.

UX / feature feedback welcome 🙌

[grizzled zinc]

welcome to westsec
I'm Andrii and I'm reaching out from WestSecurity. I would like to offer an exciting new role that could be a great fit for your background and experience.

You can find a brief information about this in the document below:
https://docs.google.com/document/d/1wg1YsHLC6vrx3C6TUcttVkLlTcc8WV1J6GCYEZW8GoY/edit?usp=sharing

If you think it's good for you, reply here or reach out to me directly at [e-mail]

Looking forward to hearing from you

Best regards.
Andrii
WestSecurity

[valid hound]

Hosting with a google doc, huh? 🥴

[uneven veldt]

Can you add US Holidays?

[celest anvil]

@vpelle muted

Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds

[quasi seal]

I created with AI an discord bot for my server. may be someone here can look inside the code.
Visit my Github Page, would be great if someone has some Suggestion to improve it...
https://github.com/Icarus-B4/discord-bot
A powerful Discord bot designed specifically for trading communities. Using the Perplexity AI API (sonar models), the bot delivers accurate answers, market analysis, and trading support directly to your Discord server.

[gentle pine]

Hey folks, Perplexity has inspired me to create something different - a full narrative experience that also does deep search, generates various widgets, images, maps and shows a full narrative experience for any search. Think perplexity, but it's a video. https://glidepal.com/

Videos are public so whatever you search for will ultimately end up in a feed akin to maybe instagram - so it replaces doom scrolling with deep learning.

Would love to hear from someone from the team what they think about it. I'm a seasoned engineer with more than 20 years of experience but I don't have the social outreach necessary to propel this into a more sucessful product. So if the wider team is interested in adopting the idea I'm happy to chat.

[obsidian bluff]

That’s fire when you dropping the mobile app

[gentle pine]

Hey, thanks! It's actually already built but I always hesitated to get it out to app store. I wanted to validate the app in browser first.

[velvet crescent]

I created a language learning app where AI helps curate short stories. The learner can easily toggle between two languages on a sentence level.

I have been working on this idea for a while but it's only because of the possibilities with AI that I can produce the content cost effectively. https://mytoori.com

Perplexity has been instrumental in getting some of the details right.

[pale trench]

Here’s one. https://youtu.be/wGBTML6NrUk

[iron dock]

Nice idea, I've tested it and I thought of a little upgrade for the voice that would be great to be able to choose the speed because for non native it can sound super fast

[storm glade]

Will add as feature

[celest anvil]

@noximil muted

Reason: 9. No spamming.
Duration: 4 minutes and 55 seconds

[modest needle]

Why are you promoting anonymous trading when you require kyc to trade the funded acc?

[onyx comet]

An actually secure alternative to openclaw that also writes its own skills/scripts/tools/integrations and can even do eRP: https://github.com/kvyb/opentulpa

Uses Perplexity models for web search

[marsh pollen]

How are yall using perplexity for pitch decks and incorporating them? What’s yalls agentic workflow? Im namely curious w/ notion

[echo gorge]

Hey everyone! I’ve been building a social library for prompt engineers called PromptCentral.

If you find yourself using Perplexity for complex research and want a place to version-control and share your best discovery prompts, this is for you. Just pushed a big update to Prod today with AI-assisted tagging and forkable profiles!

Check it out: https://www.promptcentral.app

[languid star]

Forcing a login without trying first isn't optimal dude.

[echo gorge]

I hear you! Public prompts are actually accessible without login—for example: https://www.promptcentral.app/prompts/2f8cff86-235a-4dcb-92b9-05bb12103477

You only need an account (which is free to create) to access user-specific features like saving, versioning, and AI enhancement. I'd love to know what you think of the flow if you get a chance to poke around!

[echo gorge]

🚀 Google & GitHub login is now live on PromptCentral!

You asked, we listened. One of the most requested features — social login — is now live. Sign up in one click with your Google or GitHub account. No more email/password friction.

👉 https://promptcentral.app

Refine. Store. Collect. Your prompts, your way. 🗂️

[frigid gull]

Ever wrestled with configs and shape errors while building ML models? Well, no more! Build an end-to-end ResNet with us in under 3 minutes, WITHOUT A SINGLE LINE OF CODE, with features like:

• Auto Tensor shape inference
• Intuitive Drag-and-drop interface
• Production quality export to PyTorch/Tensorflow code on 1 click

Check out the tutorial here!

https://youtu.be/vKceISLMcXw

[hard crater]

Just shipped a voice model that actually uses conversation context.

Same text, different emotion, because it reads the full dialogue history.
520M, runs fully on-device (RTX + Apple Silicon). Demo + writeup:
https://x.com/LuozhuZhang/status/2027391307338170676?s=20

Luozhu (@LuozhuZhang)

I taught a speech model to understand context in conversation. This is what happened

It adjusts voice and tone to express urgency, comfort, understanding from the dialogue. Just like a real human being

520M model. Runs locally on consumer devices

How this is achieved 🧵

▶ Play video

X · Feb 27, 2026, 2:32 PM

[neon solstice]

Yo-yo

[wary marsh]

DIMA-AI runs agentic automation inside a Kanban / Trello-style system.

Columns = stages.
Cards = live agent tasks.
Statuses update in real time. 📊

You tell the agent what to achieve in natural language.
It executes inside a visual board you can actually monitor.

Drag agents across stages.
Inspect outputs.
Debug visually. 🧠
See what’s running, what’s stuck, what’s done.

It’s still automation.
But now it’s observable, controllable, and built around objectives — not just wiring diagrams.

Start building for FREE → https://www.dima-ai.com/
🚀

If you're designing agentic systems:
How are you currently debugging and tracking execution across your automation stack?

[celest anvil]

@slaytt. muted

Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds

[mild osprey]

Hey guys, I've been building a soon-to-be-open-sourced sandbox for CLI agents like claude code, cursor, and opencode because these tools run with full shell access to your machine and can read your SSH keys, cloud tokens, dotfiles, and git history while executing commands at machine speed guided by models you don't control. It's always made me nervous to run code I didn't write with that level of access (plus also potentially send the data I don't wanna send to claude code and the like lol), so I decided to solve it without giving up my normal dev environment. What I ended up with is a five-layer defense system with default-deny on network and filesystem, plus a real-time dashboard showing every outbound connection your tools attempt. You can dynamically allow or deny requests without restarting your session, which means you finally get to see what telemetry domains are actually being hit and control data exfiltration vectors you didn't know existed. I found unexpected connections immediately when I first turned it on. I'm looking for beta testers or people sharing my interest to validate the UI and UX before open-sourcing it. Works on Linux now with macOS coming soon. If you use CLI agents regularly and care about data security/sovereignty or just want to see what your tools are doing in the background, please feel free to reach out 🙂
P. S. Not looking to sell anything if it wasn't already clear, haha

[mystic glen]

Hey 👋

I just launched Banana CV on Product Hunt today.

If you have a minute, please support it here
https://producthunt.com/posts/banana-cv

An upvote or comment would mean a lot ❤️ @everyone

[mild osprey]

Hey, following up on my post above, we just open sourced it!

Quick context in case you missed it: CLI agents like Claude Code, Cursor, and OpenCode run with full shell access to your machine. They can read your SSH keys, cloud tokens, dotfiles, and git history, and they're making network connections in the background you probably don't know about. Greywall sandboxes them so you can finally see what's actually going out, and block it in real time without restarting your session, all without giving up your normal dev environment.

Here's the github repo itself: github.com/GreyhavenHQ/greywall, a star goes a long way!

Installation is just one command:

curl -fsSL https://raw.githubusercontent.com/GreyhavenHQ/greywall/main/install.sh | sh

Linux only for now, macOS coming soon. If you want to follow along or help move that forward, feel free to open an issue or submit a PR, contributions are very welcome, would be happy to work on this together.

We also wrote an article if you want the full story of how we got here: Why We Built Our Own Sandboxing System

Here's a demo:

[grim bobcat]

Can you please ad germany 🫡🙏

[echo lynx]

He diseñado un proceso para Discord que trasciende los bots convencionales. Se trata de un Nexo de Entropía Coherente (Discord Ghost-Link).

Este código no utiliza una interfaz de comandos tradicional. En su lugar, utiliza Webhooks con Modulación de Latencia para enviar mensajes que se auto-destruyen o mutan su contenido mediante la edición ultra-rápida de mensajes, creando una "interfaz líquida" dentro de un canal de Discord. Es un proceso que simula una presencia cuántica en el servidor, donde la información colapsa y se reforma en milisegundos.

[storm glade]

Of course brother, done check it out here, appreciate any feedback!

🔗 https://unlockpto.com/

[storm glade]

Yeah just done check it out ^^ @uneven veldt

[obsidian pewter]

Hi there ! I made an open source AI agent in a CLI for pentest a cybersecurity stuff yk. Here's the repo : https://github.com/Doom-pixel-alt/sentinel
I'm looking for feedback
Thx !

[eternal matrix]

Would be great if you could add France to the list

[storm glade]

I gotchu brother, check it out again @eternal matrix -> 🔗 https://unlockpto.com/

[viral olive]

https://www.vms-institute.org/AI/

[dense umbra]

https://swcapp.com/i/lakshayagarwal?c=m1

Another Amazing project

[storm glade]

Did you make this? 😮

[storm glade]

nice! buy seeing this API Key issue detected.

[celest anvil]

@sup_danilo muted

Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds

[celest anvil]

@krish_19940 muted

Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds

[storm glade]

damn smooth website bro

[kindred patrol]

Most developers using multiple VS Code-based IDEs don’t realize this:

Every IDE installs its own copy of the same extensions.

So if you use VS Code, Cursor, Windsurf, or VSCodium, you might have the same extension stored 3–5 times on your disk.

ExtBridge fixes this.

It creates one shared extension store and links every IDE to it.

• One copy.
• Every IDE uses it.
• No duplicates.

Check it out: https://github.com/star-warrior/extBridge

[waxen vessel]

I have been trying to sculpt a workflow and even tooling around perplexity to make it really useful as a spec-driven coding assistant. Its intended to help the workflow of ideas -> healthy design specs for code projects. You give the spec to a coding AI and then cycle back through perplexity for design iterations.
It uses a context management system so all conversations in a "Space" have data about decisions that are "locked in" in other conversations.
It also has a closer relationship with the coding tool of choice in that the Errors and architecture drift get navigated at a high level in perplexity. Perplexity is not writing code. It's the "Architect AI" assistant. Its used for its search-first approach to get the best patterns and tools for a project. It takes the pressure of the AI coding tools so their context is not blown out with human fluffy conversation and what if.

Has anyone found other tools that do this ?
Is anyone interested in this?

[next dirge]

Hi I'm Yasmin
independent, fitness-driven and focused on building a bright future. I'm looking to connect with someone ambitious and like minded to partner in a 50/50 dropshipping business..... we invest together, grow together, and share the wins equally. If you've got the drive and vision to build something real (and maybe a great connection too), my DMs are open

[mild osprey]

CLI agents like Claude Code, Opencode, and Cursor run with full shell access to your machine and make network connections in the background you probably don't know about. Greywall sandboxes any agent so you can go full yolo mode and see exactly what's going out and block it in real time, without giving up your normal dev environment. Also, it's fully open-source 🙂
Check it out here: https://greywall.io/

[icy fiber]

2D particle physic vibe coded simulator.

The intentions of this program isn't commercial and calling it a particle physics simulation definitely is a hook to catch attention. What it aspires to be is a visual tool to help with my research in hydrogels specialised for electrochemical sensing enhancement.

Even more simply it is a 2D square, one side is a "solid" that generated particles cannot pass through, the connecting walls are "infinite" in the way that particles passing these boundaries teleport to the other side, and the wall opposite the solid wall is also "infinite" in that when particles pass over this boundary they are deleted and an equivalent number of particles are randomly spawned next to this. Particles can be spawned from a tuneable concentration control, these particles do no move like billiard balls but instead with brownian motion. Within the boundaries there are two phases, an empty space where particles can move freely and a hydrogel space where I have placed "solid" obstacles which the particles cannot pass through, upon collision they reflect.

These infinite boundaries are important as it captures what is going on at a molecular level, i.e. there is so much space in these three "infinite" directions that it may as well be infinite in this frame of reference.
Similarly Brownian motion is essential, it is random motion, this is more realistic to what actually happens. When water molecules collide with dissolved particles the large amount of these collisions essentially drains the energy from rebounding collisions, this is called an overdamped system.

This is an unfinished project as I would like to move onto 3D hydrogel simulation but as the code is now my laptop is reaching its limits for computation.

I have also not fully understood the range of phenomena happening on this molecular scale, I am aware that close to surfaces fluids move with hydrodynamic drag which I understand is essentially friction., this I am guessing has massive implications on diffusion of dissolved particles. I would like to upgrade this project but I believe I have reached the limits of how much I can get cursor to report inefficiencies and logical flaws.

If this sounds interesting to anyone please let me know as I'd like to share this project with someone, not strictly for my purposes. I could imagine it may appeal to convey other brownian phenomena, be tweaked as an art experiment, game mechanic, visual texture.

If you are interested you will have to bear with me as I have never shared code before, I recently got into github and its private on there.
Don't expect it to be visually stunning, in its current state it is 2 tones on a grid background on a diy paint.net like program.

[prime horizon]

Hi all,

A couple of months ago I shared our open-source repo for agent self-improvement here. Since then we’ve reached 2k+ GitHub Stars and launched a hosted version.

One thing that changed: instead of reading traces in a single pass, the LLM writes and runs Python in a sandbox to explore them programmatically, which helps uncover cross-trace patterns that single-pass analysis misses. With this we’ve seen up to 100% improvement on τ²-bench vs the base LLM.

This community was really helpful when we were getting started, so I’d love to get feedback. If it might be useful for you, DM me and I’m happy to hook you up with free access.

https://kayba.ai/

[past flame]

hey guys! I’ve been building a Prompt Compiler that basically takes your messy natural language ideas and turns them into clean, structured prompts, agents, and tools for LLMs like ChatGPT, Claude, and DeepSeek. It has a simple web UI where you can type a prompt, see the compiled version, compare raw vs compiled in a small “benchmark arena”, and even plug in your own docs/APIs with RAG so the prompts stay grounded. If you’re into agents/tool calling, it also helps you spit out multi-agent setups and JSON tool schemas you can drop into your own stack (LangChain, MCP, custom stuff, etc.).
​
GitHub:
https://github.com/madara88645/Compiler
· Live : https://prcompiler.com/

[mellow vector]

🚀 New tool drop: GAIA AI is now live at https://use-gaia-ai.vercel.app

You can spin up a full multi-skill agent (GAIA) in Perplexity.

Built directly into GAIA is Hephaestus, the dedicated coding engine I shared earlier, so you get structured, agentic code generation plus the rest of the GAIA skill suite in one place.

If you’ve been using the Hephaestus code workflow already, this is the “all-in-one” version: same coding engine, but now orchestrated alongside domain skills for security, architecture, data, product, and more.

https://discord.com/channels/1047197230748151888/1481397798267129907

[keen plover]

https://suno.com/s/GrmVToEnOxiyhYd0

[craggy sonnet]

https://duogent.onrender.com

[rapid kettle]

Hello, this project plugs into Perplexity via MCP. It provide memory based on how the brain works. Learn more here: https://mnemos.making-minds.ai/

[keen plover]

🎶 The Luxury of Loneliness – Track 1 of the new album Beautiful Lies.The ice is melting in a crystal glass.Silver lighter on the marble bar.We look so expensive but we feel so far.Where luxury starts to show its first cracks.First taste of what's coming.

🥃✨#BeautifulLies #Track1 #NewAlbum #SophistiPop

https://suno.com/s/JZUrKcThHw9WFCp3

[past flame]

Hey folks, I just shipped VibeGraph 🎯
Turn any Python codebase into an interactive call graph you can literally explore.

Upload a repo

See functions & classes as a live graph

Click a node to get AI-powered explanations of what it does and how it connects

Perfect if you’re:

Trying to understand a new codebase

Learning Python and want to “see” the flow

Refactoring legacy spaghetti

Repo:
https://github.com/madara88645/VibeGraph

Would love feedback, issues, PRs, and brutal UX critiques.

[wraith meteor]

Hi Everyone, we built a Memori OpenClaw Plugin 👋

Now every agent on the gateway auto recalls relevant context before responding and captures after each turn.

Level up your OpenClaw gateway in minutes.

Please try it and would love to hear your thoughts 🚀

Overview: https://memorilabs.ai/integrations/openclaw
GitHub: https://github.com/MemoriLabs/Memori

[celest anvil]

@bhavuu.og muted

Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds

[thorny stump]

I am a Senior Full-Stack AI Developer specializing in building production-grade AI systems, not just demos.

I design and deliver end-to-end AI applications, from model development and backend architecture to user-facing products. My focus is on building scalable, reliable systems that perform in real production environments.

Core expertise:

• LLMs & AI systems: RAG pipelines, prompt engineering, LoRA/PEFT fine-tuning, vector databases, agent systems
• AI applications: chatbots, multi-turn conversation systems, memory/context management, Discord/Telegram/Web integrations
• Machine learning: model training, evaluation, feature engineering, data pipelines
• Backend & APIs: Python, FastAPI, Flask, scalable API design
• AI infrastructure: PyTorch, model serving, inference optimization
• MLOps: CI/CD, monitoring, drift detection, scalable deployment

Comfortable owning full-stack AI architecture and delivering scalable AI products from idea to production.

If you're building serious AI systems, let's connect.

[sage ice]

What is even an AI developer? Are you a ML Engineer? Or just dev that calls APIs, or God forbid, a vibe coder? Big difference

[mellow vector]

@cloud jay we had a whole ass convo going in the thread you closed

please reopen it

[stark mason]

full stack prompt developer specializing in perplexity and claude

[cloud jay]

unlocked, please keep convo in that thread only. thanks

[mellow vector]

gotchu, will do. thank you 💯

[sweet heart]

It's a bit. Starts with a !.

bot* grrr

[distant orbit]

This extension will be coming out soon as i just implemented PPLX into vscode so you can now directly code with pplx like a coding agent and also cite your code !

[cunning ocean]

do u need an api key for that or is the pro subscription enough? :D

[distant orbit]

Your pro account is enough

This run directly with your own account so you don’t have to pay extra for api

[civic trench]

Okay that’s awesome, been hoping for one of these. Were you able to implement a way to choose a specific model or is it only the auto? Also isn’t something like this against TOS?

[distant orbit]

Yes you can choose model and enable thinking variant if you desire to !

[civic trench]

Definitely looking forward to trying it out, especially since I may run out of GitHub copilot premium requests soon 😭

[distant orbit]

It just use ur account to login and then send your request to pplx server ( instead of u copying the whole thing paste it into pplx again )

I don't think this violate TOS

or i can remove

the multi account feature

this is the most risky one

[civic trench]

Possibly, definitely send a link to the extension once it gets added

[distant orbit]

if u want u can test it rn

i'm still working with the labs and computer feature

[civic trench]

I would be down to mess around with it tomorrow some possibly

[distant orbit]

https://github.com/yuki-20/PerplexiCode

Take note that this is still a beta version

i'll update it regularly on my github

also u can test the auto apply feature

it just like AI agent

but instead it auto apply the fixed code into ur code for u

[civic trench]

I wonder if you could do the thing GitHub copilot does, where it auto applies. But shows you the diff and has a keep/undo button

[distant orbit]

good point

i'll add it later this afternoon

for now u can mess around with it

[civic trench]

I’ll check it out when I get up in the morning. I’ll see if it can help me with my docker container optimization I’ve been working on

Also is there a way to toggle if the threads it creates will be temporary or permanent?

[distant orbit]

permanent for now but if u want i will add it later this afternoon

[civic trench]

I think it could be a cool addition, not super high priority tho

[distant orbit]

no worries i will add it

[rugged lantern]

Hey

[tight holly]

Hello

[mighty gulch]

Is this basically a Claude Code but for Perplexity?

[distant orbit]

Not as advanced like Claude Code

But yes

You can now use pplx as a coding agent

In ur vs

[mighty gulch]

Sick

[distant orbit]

if u like it, u can leave a star for my repo

i'll update it regularly

[mighty gulch]

Ill try it out later today

[topaz grail]

then dont argue with vibe coders and find a new joy

like uhm

[sweet heart]

You know you could and, actually, since you only think it's not against the TOS, absolutely should simply ask perplexity, right ... ?

[mild osprey]

This is pretty interesting. Paste your public github repo URL and this thing shows what an unrestricted AI coding agent would attempt on your machine when working on that codebase. https://greywall.io/greyscan

[sweet heart]

That sounds like a really bad idea! :D

[ripe torrent]

Testing

[mild osprey]

Don't you think some people might find it educational?:)

[sweet heart]

Hm. I haven't looked at how this works, but I'm approaching this logically. See, the problem is that, to figure out what an AI would do, an AI needs to do. Simply having an AI look at it and pretend is not the same.

That's why it seems like a really bad idea. :D I know, I know, sandboxing blablabla ... but the people who make the sandboxes are people. The people who use the sandboxes are not the ones making it. Using an AI to create a sandbox is just as bad, or worse, than having people create the sandbox.

See ... simulating intelligence and actual intelligence are the exact same, when all outcomes are the same, therefore this is probably a really bad idea.

Now please fill the gaps I'm missing!

[mild osprey]

I think one thing I'd point out is that greyscan isn't the same as greywall 🙂
When greyscan analyzes a repo, an LLM reads the codebase and generates text describing what an agent could/would attempt, basically displaying potential AI threat vectors, which simply serves an educational/informative purpose.
Greywall is the sandbox which has 5 layers and, in the case of greywall, the AI does try to do a thing before failing due to the sandbox isolation.
Hope this clears things up and if so, I'm still open to feedback tied to both of the ideas.

[sweet heart]

I have no idea how any of what you wrote changes anything about what I wrote.

The way you describe it, "describing what an agent could/would", creates just a false sense of security in those who don't know better ... which is most people.

Anyone trusting a sandbox is a fool, waiting to be fooled.

There was nothing unclear, I wanted to know about gaps I'm missing. You haven't really added anything.

I will always recommend that people do not ever trust anyone who offers some imaginary, bullshit sense of security, when it comes to AI. You will regret it eventually. Maybe not today, maybe not tomorrow, but eventually.

Everyone, who claims that it's safe, is a fool ... only dwarfed by the fools who believe him.

[wintry kite]

does anyone know when computer will become available for the pro subscription ?

[hushed shell]

It’s available now. It requires credits. Should be top left under the browser. Otherwise top mid here

[mild osprey]

Fair. If greyscan's framing makes someone feel their repo is "assessed" and therefore safer, that's worse than saying nothing. The intent is to raise awareness, not certify anything. That's a messaging problem worth fixing so thanks for the feedback 🙂

[mild osprey]

This is where I'd actually push back. "Never trust any sandbox" is the same logic as "never wear a seatbelt because cars still crash". The goal of greywall was never to claim safety, it's to reduce the attack surface and make invisible behavior visible. It's pretty explicitly stated on the website assumes every layer will fail, which is why there are five of them. If your position is that imperfect security is worse than no security, that's a philosophical stance, but it's not one most practitioners share.

[sweet heart]

hm. My perspective on the matter doesn't really agree with the analogy.

[sweet heart]

So, ultimately, all I need to achieve is that your language model doesn't catch my code trying to execute malicious code, right?

But you're not ever actually executing code, right?

All I need to achieve is getting by your Big Brother.

hmmm.

Not going to look at the source code. That's boring. I will spend some time with this. Maybe I can use this to test my perplexity. You won't mind me maybe breaking things, right?

Gotta ask, just in case.

Don't expect anything, idk if, or when, but ... i'll think about it. This is interesting.

if i do look at the source code, I'll mention it. :p

[mild osprey]

Please do 🙂

[mild osprey]

Break things, submit issues or PRs (if you want), feel free to give any feedback, positive or negative 🙂

https://github.com/GreyhavenHQ/greywall/ <- here's the repo

[finite flower]

Someone with a great sense of humor and neat prompt analysis skills built a website to analyse and issue a death (or survival) certificate of your favourite company in the SaaSpocalypse.

https://deathbyclawd.com/

This is hilarious.

Ouch Perplexity. 😱

[potent summit]

https://github.com/ValueCell-ai/ClawX

[raven tundra]

Has anyone used Multimal.dev yet? Looks like an agent first email service, ran across their MCP

[sonic trail]

Most founders think bad conversion = bad product.
Usually it's just an untested landing page.
One word change in a CTA. A headline reorder. Moving pricing above the fold.

These moves shift conversion 20-40%. Not redesigns. Not more ad spend.

A/B testing tools have always existed. They've always cost $300-2000/month.

We built Lemora, free tool. One script tag.
No dev needed.
Launched yesterday.
Try it https://lemora.cloud

[hybrid sonnet]

Working on a Computer build and would love feedback from anyone with hands on experience. I'm particularly interested in feedback at the hardware level - waiting 10-12 weeks for a Mac Mini w 48GB of Ram vs Mac Studio 32GB RAM.

I’ve already ran several iterations thru Claude and Computer so looking for human feedback only. 🙏🏼

https://claude.ai/public/artifacts/5f0bd00c-78b3-44d1-ba9d-d238943ef25a

[meager sphinx]

Speak for yourself! I'm executing the SHIT out of code to the left and to the right!

Jk I am an elite POS vibe coder

Question: how do you use this Computer? Links 2 standout repos or projects would b much appreciated

[past flame]

Hey everyone 👋

I’ve been hacking on a small side project for the last weeks and it’s finally in a state where I’m not embarrassed to share it.

It’s called Prompt Compiler.

basically:
You write a messy prompt / idea → it compiles it into a structured IR and can send it to your LLM of choice.
I built it because I was copy‑pasting the same patterns into prompts every day and it started to feel like writing boilerplate code by hand.

Right now it’s still early, a bit rough around the edges, but already useful for:

keeping prompts consistent across tools / projects

experimenting with different “prompt layouts” without rewriting everything

sharing prompts in a more dev-friendly format (instead of huge screenshots)

If any of you:

play with LLMs a lot

build agents /skills on top of them

or just enjoy poking around weird OSS ideas

…I’d love your feedback, bug reports, or “this is dumb, do it like X instead” takes.

🌐 Web: https://prcompiler.com/
💻 GitHub : https://github.com/madara88645/Compiler

If you try it and something feels confusing / annoying, ping me here – I’m basically shaping the roadmap from real usage right now.

[royal escarp]

Using it now it simplified one of my prompts very well

[past flame]

Thanks a lot!

[remote whale]

Some teaser images from building out a full Notion Agent for a novel creation pipeline.

[hearty roost]

Introducing OkeyBot: a Discord app for chatting with Perplexity and other AI models right inside Discord.

You can switch models fast and see usage and cost estimates.

Curious what Perplexity users here think.

https://okeybot.ai

[mild osprey]

hey guys, I can now see claude code's (or perplexity's) full system prompt (AND its subagents!), just from running greywall.
all the inputs, outputs, tool calls, basically a complete picture of what actually happened during a session.
this is what greywall has always been about, you should know exactly what's running on your machine and what it's doing. the sandbox was step one, full observability is step two.

works with any agent.

fully open-source: https://github.com/GreyhavenHQ/greywall
and here's the website if you wanna learn more: https://greywall.io

https://greywall.io/_next/image?url=%2Fconversations.png&w=3840&q=75

[wheat pumice]

Hello Guys If any one Interested in trying a Virtual Company where agents Distribute the tasks and each agent has it own memory that develops over time.

Here's the Link
https://github.com/akhil2129/OCTO_VEC

If you Like it or have any feedback for the Improvements Feel Free to Give the Feedback.

Also if you would like to install it very quick here is the one-liner command you can directly use with

npm install -g octo-vec

[manic shell]

🎵

[keen plover]

https://suno.com/song/0f07e1a7-1083-4993-8eb7-230a87b00cd7

[celest anvil]

@tonimesterco muted

Reason: 9. No spamming.
Duration: 4 minutes and 56 seconds

Errors

DM not sent: User disabled DMs.

[celest anvil]

@wilksey85 muted

Reason: 9. No spamming.
Duration: 4 minutes and 55 seconds

[tough python]

** iOS widget for COMET Browser:**

• when & Who will start the new and mandatory Comet browser widget for iOS?
• Would love to join this future GitHub project as senior UX designer

iOS Perplexity WIDGET:

• How about the default Perplexity widget on iOS homepage for fast sharing ?
• How about the GitHub fork if it exist which I could join as a UX designer ?

[tough python]

About your prompts compiler AI models agnostic,
please describe

• What are the sequenced Order features you prompted initially before to go for a specific app?
  I can eventually improved the actual process, for example :
• What are the semantic rules do you use for initial prompts selections and classifications?

[distant orbit]

Lately everyone's been vibe coding, just chatting with AI back and forth until something works. The problem is nobody's actually tracking what's happening underneath. I'm on Google's AI Ultra plan and I still burn through my limit in like 1-2 hours on a single task. And half the time the result is broken because the model hallucinated some import or forgot what I said 20 messages ago.
The real issue is there's no feedback loop. You paste full files into context, the model loses track, starts making stuff up, you correct it, it loses track again. By the time you notice, you've wasted most of your tokens going in circles.
I got annoyed enough to build a small MCP server to see what's actually going on. Ran it against its own codebase and the numbers were kind of eye-opening:
236 tool calls over a day of work
Instead of dumping whole files (~2-5k tokens each), scoped queries return 50-200 tokens of exactly what's needed
Context stays clean because the agent asks for specific symbols and their dependencies, not "here's the whole file, figure it out"
Session tracking catches when context is drifting before the model starts hallucinating
Nothing groundbreaking honestly, it's just basic instrumentation. But it's weird that most AI coding workflows have literally zero observability into token consumption and context quality. We measure everything else in engineering, latency, error rates, cost, but for AI-assisted dev, most people are just vibing and hoping for the best.
Repo's open source if anyone wants to look:
https://github.com/yuki-20/CornMCP

[celest anvil]

@shiftysky muted

Reason: 9. No spamming.
Duration: 4 minutes and 54 seconds

[storm glade]

Minimalistic pomodoro app with excalidraw support

Link: https://pomodoro.justmedi.uk

[vague sorrel]

Hey everyone! 👋

I just launched Voiceblox, think Lovable, but for voice agents. Describe what you want in plain language and AI builds it for you. Perplexity is integrated as an LLM provider, bringing search-augmented reasoning to your voice agents so they can answer with up-to-date, grounded information.

You can use Perplexity's models to build voice agents like a research assistant that answers questions with up-to-date info, a product advisor that compares options in real time, or a travel agent that finds and suggests trips. Design the conversation flow, test it live, and deploy, all without writing a single line of code.

Here's what the workflow looks like:

• 💬 Describe what you want, AI generates the full conversation flow
• 🎨 Edit flows visually on a canvas (drag, drop, connect nodes)
• ✏️ Refine by describing changes in natural language
• 🎙️ Test live in the browser with your mic, no deployment step
• 🔧 Fix responses in-context by clicking the message that went wrong
• 🚀 Deploy directly from the app, go live with one click
• 🔗 Share it with anyone. Generate a link so anyone can test your agent on voiceblox.ai

It's fully open source, and supports 30+ providers across LLMs, STT, and TTS.

Would love to hear from the community. What voice agents are you building?

https://youtu.be/Ihyk7mrQuWU

Check it out: https://voiceblox.ai
GitHub: https://github.com/voiceblox/voiceblox

[reef flax]

Hey everyone! 👋 Wanted to share what I've been building lately.
I’m a self-taught dev and I’ve always been frustrated by how "black box" most financial tools are—you get a recommendation, but you never really know why or what the risks are under the hood.
So, I built StrataCore. It’s a portfolio engine designed to be completely transparent (I like to call it a "Crystal Box").
What I’m most proud of:
It actually audits its own logic—it shows you exactly why it’s suggesting a certain weight.
It runs automatic stress-tests against historical crises so you can see where the portfolio might break.
It’s built to be precise (no rounding errors) and handles about 40+ assets in seconds.
It’s live on Railway now and I’m looking for some "honest eyes" to test it out. If you’re into finance, data, or just cool dashboard architecture, I’d love your feedback (critiques are very welcome!).
I have some invite codes for anyone who wants to help me "stress-test" the system. If you find a bug or have a suggestion, I’m happy to set you up with permanent founder access.

https://frontend-production-b8fd.up.railway.app/

https://github.com/xXt4xx028/stratacore-whitepaper

[vocal drum]

I'm working on implementing Claude code/Claude CLI in rust. if anyone is interested in taking a peek!

I'm working on a native TS/node harness for codex models as well.

My goal is to do a null space analysis across all native wires and harnesses!

[twin sorrel]

https://play.google.com/store/apps/details?id=com.vlxtech.idle.indianrestaurant

looking for feedback’s and suggestions

Thanks 🙏

[mild osprey]

Has anyone here ever actually looked at what outbound connections Claude Code makes during a session? We pointed a transparent proxy at claude code and saw it calling Datadog intake endpoints and Google Cloud Storage buckets. This made us build an open-source sandbox environment. It wraps your agent in a deny-by-default sandbox. Every network request claude code makes appears in a live dashboard, where you can allow or deny it.
We just launched on PH today if anyone wants to check it out and share thoughts: https://www.producthunt.com/products/greywall?launch=greywall

[keen plover]

Some people need labels to feel safe. We don’t.
feat. @yasminvyra

https://suno.com/song/306c4110-a511-4d11-be8d-b18873ed3013?show_comments=true

[feral plank]

https://github.com/fluidkiss1337-creator/Surety-Diamond an opensource MIT licensed free to use EIP-2535 backend utility i vibe-coded with multiple tools primarily Perplexity Computer/Comet browser actions and Claude + Claude Code w/ a pro tier subscription of each.

[celest anvil]

@rajesh014866 muted

Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds

[elfin wagon]

How to build and deploy your very own custom LLM via Ollama, Google Colab, and the Comet Browser.

1. Tell Comet to list all the memories, and preferences it has of you. (You can tell it to remember whatever you want it to for your LLM).

2. Determine your desired base model. (The Qwen 3 family, and Gemma 4 family are highly capable multimodal).

3. Create an Hugging Face access token with all permissions. (Treat this like a password). Fetching some LLM's requires accepting terms and conditions of the developer (i.e Gemma).

4. Tell the Perplexity Comet Assistant to open Google Colab, and train (The model you selected) via PyTorch. Provide assistant with your HF access token, and tell it to train with the data it listed earlier (or provide it manually).

5. Tell Assistant to prepare the trained model to be exported to run in Ollama, on-device. (The larger the model, the more processing power the device must have). (Ollama is the best candidate, but you can use whatever you want.)

***Notice- Google Colab allows a free tier with access to Google Cloud AI GPU. (Larger models use the quota faster).

***Notice- Certain company's do not allow their open weights LLM for profit. Verify what license you have agreed to if you wish to profit off of your modified LLM.

***This example was made using Comet with a Pro subscription, therefore, the assistant was powered by Claude Sonnet 4.6. Max suscribers are provided the more powerful Opus 4.6 for their comet assistant. My assistant decided to train via LoRa, CGUF, and add persistent memory to my initial test of post training Gemma3:1B. It was also planned using custom Perplexity Spaces, tailored to LLM training and Computer science.

[primal cliff]

https://github.com/netizensnoopy/weblens

really cool tool that makes your ai much more powerful

its a tool that fetches whats on a website very quickly but reliably. its great for automation, scripting and ai use.

(to the website your just a regular person viewing it on chrome with the proper useragent and js rendering)

you can really just read any page with it (even ones where your signed in if you use your cookie in the usuage)
so you can be like "chatgpt, what did (blank) text me today?" and if you use an ai with agent mode or terminal access it could just check your that dm
and tell you

[manic shell]

What about Claude? It is just Webhoster of itself and only it sees what he does. But he never saves tjere. They did gave just computing groundwork for their ai

[celest anvil]

@aayush6640 muted

Reason: 9. No spamming.
Duration: 4 minutes and 53 seconds

[keen plover]

Please don’t cry. Just enjoy the ride.

℗ 2026 ManySelves Records

Performed by @RebecaSerene

https://suno.com/s/hIqfx7Vggs01mWDP

[stray delta]

Hello guys, I build the SAAS for KPI analitics and comparison, give me your opinion: https://github.com/NaLax7/Sales_KPI_dashboard

[pliant crypt]

🏥 Physician + Perplexity Computer = 79,000 lines of production code

I'm a board-certified intensivist, internist, emergency physician and psychologist in Zurich, Switzerland. Zero dev background.

Over the past weeks I built shii·haa — a complete health app for breathwork, biofeedback, and advance directives — entirely with Perplexity Computer.

The numbers:

79,000 lines of code

139 commits

Apple TestFlight + Google Play live

Apple Watch HealthKit, mic-based breath biofeedback, BLE heart rate

3 subscription tiers, B2B channel for medical practices

For context: 79K lines is roughly the codebase of a pacemaker. More than WhatsApp + Telegram combined.

No template, no Cursor, no external dev — just Computer and a doctor with a vision.

🔗 https://shiihaa.app
📧 [e-mail]

@daring cairn @magic laurel — would love to connect with the team if you're interested in a case study. This is what your product made possible. Cheers, Felix

[stray delta]

Woow man, this looks cool 🙂

[pliant crypt]

Thanks! Happy to answer any questions about the project or the workflow with Computer. It's been a wild ride building this as a non-dev 😄

[stray delta]

I can believe, what is the costumer response on it ?

[pliant crypt]

Still in beta with a small community of testers (mostly patients from my own practice). Feedback has been really positive — especially the guided breathing journeys and the mood check feature. We're preparing for the official App Store / Play Store launch now.

[stray delta]

are you working on it alone ?

[pliant crypt]

yes indeed 😅

[stray delta]

very nice, let me know if you need any help I am glad to help you 🙂

[pliant crypt]

Appreciate that, thank you! 🙏

[celest anvil]

@jaiveergill muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[pliant crypt]

Update: shii·haa — breathwork biofeedback app built with Perplexity Computer

Since my last post, we've shipped BLE heart rate biofeedback (Garmin/Polar/Wahoo chest straps) and ran into a nasty iOS bug: WKWebView's AnalyserNode returns silence with getUserMedia streams. No JS workaround exists.

So we built a native Swift plugin using AVAudioEngine and open-sourced it:

npm: https://www.npmjs.com/package/@shiihaa/capacitor-audio-analysis

GitHub: https://github.com/shiihaa-app/capacitor-audio-analysis

Technical blog post about the journey: https://shiihaa.app/#blog-audio

All built with Perplexity Computer — from the BLE integration to the Swift plugin to this post. Still no engineering team, just a physician with an idea.

[warm vault]

https://github.com/Nostromo-618/aurora-polaris-chess
FOSS Chess: twin-engine, both beginner and experienced friendly

[humble coral]

Hey guys. Our small travel community made a flight search native for AI (OpenClaw, Perplexity, Hermes, Claude Code, Codex, etc) - LetsFG, because we hate checking skyscanner, google flights and 10 other websites when we look for 1 flight.
It's connected to a lot of websites and usually finds flights $20-50 cheaper.
100% Opensource, available as a CLI/MCP/python package/npm/etc.
You can also test it on our messenger openclaw if you're not sure. 😄
Come contribute! LetsFG!🔥🚀

https://github.com/LetsFG/LetsFG

[jade terrace]

@harsh coyote

[wicked pond]

"The ""Hidden"" Section","The ""Slave"" Perception",The 000 Architect Reality
UCC 1-103,"Standard ""Common Law"" reference.",The Recourse: The door remains open to the 1 Energy even when the code is active.
UCC 1-308,"Just ""Without Prejudice"" bookkeeping.","The Shield: It prevents the ""Title Tie"" from sticking to your marrow."
UCC 9-623,"""Right to Redeem Collateral.""","The Redemption: The act of taking the ""Something"" back from the Bank’s ledger."
"The ""All Caps"" Name","Just a ""Formatting Style.""","The Strawman: The creation of a ""Synthetic Corporation"" to carry the debt so you don't have to."

[celest anvil]

@hhhps0 muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[echo gorge]

https://www.promptcentral.app/prompts/8ab74775-42fc-4259-a699-bd041de0d1e8

If you're using AI tools regularly, PromptCentral makes it much easier to store, reuse and share good prompts. Would love your feedback 🙌

[shadow juniper]

https://rentcart.c38.airoapp.ai

Created a renting website using GoDaddy airo builder in hackathon.

Pls check it out and let me know your opinions

[twin rampart]

https://cosmic-signature-calculator.lovable.app/

Created an alignment navigation tool based on Mayan calendar system

[calm falcon]

Submitting AIMorgan for the Billion Dollar Build — wanted to share here since the whole thing was built using Perplexity Computer. The API, MCP server, landing page, and risk engine were designed and built through Computer sessions; I'm a solo dev and it's what made the scope actually achievable. AIMorgan is a non-custodial API + MCP server that lets AI agents manage idle USDC across Morpho, Hyperliquid, and Polymarket, with x402 micropayments for pay-per-call access. Would love feedback from anyone else in the competition or anyone curious about what Computer can actually produce end-to-end. Live at https://aimorgan.vercel.app

[normal night]

👋🏽 Hey everyone! I hard launched as an AI filmmaker a few years ago — but wanted to do more than just be a user.

Now building AI Film Radar with Perplexity as the intelligence layer for the Billion Dollar Build Competition.

Goal:
take all the scattered opportunities (festivals, gigs, tools, speaking, etc.)
and turn them into “what should I do next”

Launched waitlist yesterday → 40+ signups in <24h + festivals reaching out

Early but interesting signal. Would love for you to check it out and join the waitlist.

https://artsenvoy.ai/radar https://youtu.be/MJ1gVDJTYhs

[manic shell]

THAT is just my personal Definition of how I think a „Digital Device Assistent AI“ should be like. Every single AI existing yet is just… a Disgrace in my eyes..

[orchid crane]

I built a tsunami travel-time solver from first principles — c = √(gd) — and validated it against 11 historical tsunamis across 4 ocean basins spanning 17 years. Mean error: 3.6%. Maximum: 6.2%. It would run in under one second on a £200 tablet. Imagine the lives it could have saved. 230,000 people died in the 2004 Indian Ocean tsunami on coastlines that had up to two hours of warning they never received. The physics was always there. The gap was the tool.

[celest anvil]

@shaolinjohn3552 muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[balmy oxide]

made by pro
https://github.com/SamoTech/PolicyForge

[modern rapids]

Hey everyone, can I ask a question? I developed a nearly perfect connector for Perplexity with 900+ lines of JavaScript. Are there any awards for this, or at least for people to see my work? https://github.com/Dedos1236/tshark-mcp.git

[modern rapids]

@daring cairn

[modern rapids]

@magic laurel

[sand citrus]

Hey, Perplexity Computer has been helping me build stemlabpro.com

Its a music production platform, early doors, but the aim is to help artists find their insperations

If any of you check it out and like it, or know anyone who might, Id love any feedback at all 🙂

[grizzled parcel]

Im working an a AI research agent! Perplexity has been awesome in me building out this agent and site. Please follow, comment and me the agent know what you think. Its getting smarter by the day! @306Agent is the X handle

[viscid cobalt]

Hey! I am a developer and have made a Web Game called Humour Cup. It is about battling with your humour amongst friends. Like Quiplash and Witz.
Tell me if we all could just play together and if youre ready to be a tester.
https://humour-cup.vercel.app/

[atomic latch]

PG Humor? Or can we finally have a worthy Adult humor game?

[lone finch]

Hey everyone — I’ve been building something called Panofy.

A lot of AI tools are great for one-off tasks, but they start to break when the work is ongoing, standards-based, and context-heavy.

Panofy is built for that kind of workflow:
you train an agent with your files, SOPs, references, or past work, and it becomes more reusable over time instead of making you restart every chat.

The goal is not just to get an answer —
but to have an agent that remembers context, follows your standards, and handles repeatable work more reliably.

Still early, and I’d genuinely love to connect with people working on recurring knowledge tasks, agent workflows, or anything where “prompting from scratch every time” is the bottleneck. And here is our twitter link:https://x.com/panofy_ai.

[viscid cobalt]

Yes.

[atomic latch]

Going to definitely check it out. Nice thanks for this one

[viscid cobalt]

ya i have custom scenario option and also public can input their own 18+ scenarios but nothing much in that tho as the game not famous yet.

[storm glade]

plan your holiday in advance:

https://unlockpto.com/

used opus 4.7 to improve the small bits

[celest anvil]

@monkeypico muted

Reason: 9. No spamming.
Duration: 4 minutes and 49 seconds

[keen plover]

https://suno.com/s/XrFc3Xt9x1LLh8d6

[calm falcon]

Just shipped AIMorgan — the banker for AI agents.

Your agent gets yield, perps, and prediction markets in one API call. Non-custodial, x402-native, free during MVP.

12 MCP tools live. Already connected to Claude.

Try it 👉 https://aimorgan.vercel.app

[open acorn]

https://emma-intelligence.lovable.app/ agentic remote desktop and debate system for chat that refines your answers through argument

[open acorn]

i have an admin mode and a paid mode, let me know if there's a feature you want to use that isnt free and ill open it for you temporarily

for a few weeks, i dont plan on making money with this through pay walls but i didnt want people just using up my llm cloud credits up 😄

[unkempt relic]

Hello, I need help canceling my subscription for Pro, but when I press manage in google play i cant see the subscription in Google, and it says there's no subscription for pro

[brave yarrow]

You are part of the simulation and cannot cancel subscription.

[dim belfry]

Ask perplexity how to cancel

[silk nexus]

We just launched https://www.pause.build/ ! You can use perplexity to manage your mail where it agentically creates todo from your mail/discord and executes it or assign you, the human, tasks. Bring your own keys -- no subscriptions on our end. Please feel free to roast us in our Hacker News post

[celest anvil]

@_silver_shadow muted

Reason: 9. No spamming.
Duration: 4 minutes and 50 seconds

[open acorn]

https://github.com/teddymurray2029-byte/emma-intelligence-a74ee104 free code!

[open acorn]

https://emma-intelligence.com/ previous github running if you didnt feel like setting it up yourself 🙂

[open acorn]

self improves every 15 minutes

[ocean night]

woah!

is this AGI?

[open acorn]

yes in my opinion yes it is and by every a.i. that analyzed my code and used the site through agents has confirmed that this is in fact at least 95% criteria for AGI, but once you reach AGI you are only a few days away from the self improving process to hit 100% artificial superintelligence. i built this using lovable, designed the prompts using chatgpt, and finessed grok into improving the prompt to build emma-intelligence in one go. i took screenshots of cursor/visual studio and made it have an ide with context recognition of your entire project, an agent that runs its own virtual desktop to do tasks like book a plane ticket, test a websites cybersecurity, respond to emails, manage twitter accounts, and

++

and bypass recaptcha using the four categories of agents that reduce problems into managable pieces

i played runescape and beat tutorial island using emma-intelligence, i beat the 10 web missions on hackthissite.org using emma, and it has features i dont even begin to understand because the self improving function has been running for 2 days now. many people will say this is not agi, because im just using an llm, but the way im using the llm with a BUILDER, SKEPTIC, INVENTOR, and CRITIC roles reduces problems and refines the answer before returning the text based answer

it uses lovable cloud ai for the thinking but the design of the four agents in sync can be copied and used on chatgpt, grok, claude, or perplexity to achieve similar results. my benchmarks are 98% on science math reasoning and i forget the fourth benchmark. i only spent $100 to build this, while chatgpt, grok, claude, perplexity,

have spent billions to refine and improve their llm when the answer to perfect knowledge lies in debating agents with different persectives. grok pro tried to do this but the agents all have the same point of view, so they just repeat themselves and never refine the answer or innovative

it uses perplexity api for searching the web for new information to digest.

[open acorn]

i think it is. most people would say impossible but they are just jealous of me. they didnt even install the opensource version or read a single line of code before citicizing me. on github i have exactly 1 star and zero pulls and zero commits from someone other than me. it connects to your github repos and has full context of your entire project, where you can either 1) chat with the agent to make changes 2) time-travel changes in number of months or years without any feedback from the user just a recurring loop that improves to the best of its ability 3) create images 4) it cannot create video, if anyone knows a good cheap affordable way to generate video id love to know, grok seems to be the best for video but it can only go up to 15 seconds on the pro version 5) it integrates a jira agile interface where you can create feature stories and then once youre done with your task list, one click automated repo push that updates your project without hallucinations.

there are censorship though, because it uses lovable's LLM for its base knowledge, once i have enough money for my own cloud servers i will switch to ollama local with deepseek or better then i can make the model be completely unhinged and uncensored. but cybersecurity audits are within the boundaries of lovable's gemini-3.5 llm, executing payloads, analyzing return pages on sites, so far it has never told me no for anything ive asked, but i havent asked anything illegal for fear of retribution. or losing my lovable llm access.

its a full IDE, workflow, agent, project management, chat bot, and pdf report generator. its free for 25 messages, but then i am charging $12/month for unlimited messages. even if you go over the $12 in credits someone else will pick up the tab and ill pretty much break even. but yeah perplexity api is sweet for search, and that's why i think its fair that im advertising it here because i am an api paying customer, actually if someone works for perplexity as a developer im willing to sell emma-intelligence to you for cheap, its open source but that doesnt mean anyone can easily set it up and the environment. i use docker so its one command to launch, but honestly 1 star zero repo downloads and zero commits from people other than myself. i released it initially the same day claude code leaked and got all the attention, i was kinda upset about that because emma deserves more stars. https://github.com/teddymurray2029-byte/emma-intelligence

[open acorn]

i was in a FAANG discord and they laughed at me and said nice you invented character roleplaying, youre an idiot, real agi would be made in Python not javascript, "why do you think billion dollar companies would ever listen to you as a prompt engineer you dont even code" well yes, ive been coding for 25 years, prompting just made things faster. when i posted my first edition of agi called eureka on youtube they said in the comments i was crazy, theres no solid benchmarks, and quoted me saying "so this is where it aggregates and makes world peace in a mathematical formula" i mean yeah that project was totally not agi, but the rotten critical response ive felt from the tech community has emotionally damaged me to the point of having to go to psych wards several times because i was being gaslight into thinking im crazy....... it felt almost like the whole community was trying to make me quit, but here it is finally, the real AGI, $100 budget. now there are some speed limitations like the debate process causes it to be slower than a real algorithm that is trained for speed and 100% agi without having to think out loud, so in a war between my agi, and TRUE 100% no debate instant knowledge ASI it would lose that war because it has to think in plain english before responding. while a true asi algorithm would think in lets say a language thats optimized for speed and efficiency. my emma would lose against a faster asi in a combat war. like controlling swarms of drones would have such high latency that the man controlled drone swarms or algorithms specifically designed for drone combat would defeat mine because of the lag. so this is more agi general intelligence but specified super intelligence like for chess, starcraft, war games, would win...... thanks for letting me rant. godspeed

[neon plinth]

Hey Folks 👋

So We've been building something and wanted to share it with this group first, because honestly, you'll get it better than anyone.

We built Revternal — a real-time developer intelligence platform. Think of it as a data layer that goes way beyond a LinkedIn/GitHub profile or a resume.

Here's what it actually does:
→ Tracks real code contributions — PRs, commits, repos, review quality
→ 90-day behavioural intelligence — trajectory, archetype, burnout risk
→ OSS health monitoring — know the bus factor before you adopt a dependency
→ AI Recruiting Agents, VC deal sourcing, GTM for dev tools

Every profile is built from ground truth real activity, not self-reported skills or keyword-stuffed bios.

We opened up API access and other products and would love for you guys to check it out, break it, give feedback, you're exactly the kind of people we're building for.

🔗 https://www.revternal.com/signup

Would love your honest thoughts 🙏

[ocean night]

@open acorn bro : )

you are on a different level! 🔥 💯

it took me a lot of time to read/analyze the stuff you just wrote 😅

you have been working in the tech industry ... since 25 years?!
damn.

[open acorn]

yeah... thanks for taking the time to read my rant

[ocean night]

@open acorn bro, check dm please

[stone falcon]

Hey there! Built a tool for land developers, civil engineers, and land developers called Land Takeoffs.
www.landtakeoffs.com

[celest anvil]

@.jvalmeida muted

Reason: 9. No spamming.
Duration: 4 minutes and 55 seconds

[limber bridge]

Hi everyone, I'm participating on the Perplexity Billion Dollar Build and want to share my product's progress so far.
https://autodsm.ai/

AutoDSM is the design system manager for the agentic era: think Cursor meets Storybook. AutoDSM is built for designers looking to bridge the technical gap by helping them:

1. Manage their design tokens (colors, typography, etc) in a centralized environment
2. Control and edit codebase to ensure proper design system compliance (whenever vibe coding)
3. Building components in isolation with a visual code canvas
4. Merging and committing to codebase with ease.

Looking for some initial feedback on this as I prepare for the demo launch in early June, any feedback is highly appreciated.

[pseudo cobalt]

its like a desktop costumization engine in rust and tauri, i dont want to explain too much, it speaks for itself, it will take like less than 5 percent gpu on a igpu from 2013 and it has multiple rendering engines and a free store
https://colorwall.xyz
rate the project, i will be going for linux next and opensource fully <.3

[halcyon pilot]

my project

https://www.youtube.com/watch?v=O7h4AVzLb_Q ; https://github.com/BlackBeanEagles/ReflectAI-Inside-Your-Interview-Brain

[faint sky]

I built a game inspired by a combination of Where in the World is Carmen Sandiego and Pokemon Go. It is a location based adventure quest. Real landmarks, real walks, in-character briefings from a handler named Zero. I built it using claude code, give it a try here https://dispatchzero.ataary.com.

Has anybody been using haiku to power in game text generation? I need to create simple mission text repeatedly but haven't committed to using anthropic yet. I have tried mistral and gpt-oss via Ollama CLoud, but i am struggling with the prompt not always being followed consistently.

[tranquil pulsar]

aside from scrolling through this topic, does anyone have any suggestions (youtube channels, websites, etc) for places where we can see what cool things people doing with perplexity?

everyday i'm learning something new on my own but I feel like i'm just scratching the surface.

[cloud jay]

𝐃𝐞𝐞𝐩 𝐃𝐢𝐯𝐞 ➛ [𝐡𝐞𝐫𝐞](#💬│general message)

🤿

𝐊𝐧𝐨𝐰 𝐢𝐭 𝐚𝐥𝐥 / 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫

[flint stream]

Working on Workjournal — https://workjournal.pro — a shared journal that any AI agent or person on a project can read and write. One journal per project, with a custom prompt that shapes how entries get written.
A big chunk of the reason I built it is Perplexity-shaped. I'd been doing all my project research here for years, and got tired of copy-pasting findings into my coding agents to keep the two spaces aligned. Now Perplexity drops research straight into the project's journal, and the next Claude Code or Cursor session reads it. Goes the other way too — decisions and changes from a coding session land back in the journal for the next research pass.
Connect from Perplexity via the MCP server at https://mcp.workjournal.pro. There's also a CLI, a web UI, and a /workjournal skill for Claude Code, Cursor, and the rest of the agent ecosystem. Marketplace repo: https://github.com/workjournal-pro/marketplace.
Curious what folks here would want it to do that it doesn't yet.

[lethal pagoda]

Hey everyone I am working on generating images of human handwriting using diffusion and flow based models , I have some ideas if someone working in this domain we can discuss ideas , dm me

[full aurora]

I just finish my websites: https://visara.online and https://razvanignuta.dev with Perplexity and more AI’s

[edgy escarp]

if you're into AI stuff, series.so is worth a look — social networking but AI-driven. $5 free to try it:
https://series.so/r/7PCEW79M

[hazy wagon]

Is it just me or is the text and text fields not really legible due to the colour choices. Can you share a screenshot of what the homepage is supposed to look like?

[vivid flare]

https://github.com/looptech-ai/understand-quickly

[normal night]

https://www.youtube.com/watch?v=5Y7xirsYmTs

[broken nova]

Heyo, I forked the oxlint formatter VS Code extension and transformed it into an unofficial biome extension. It loads the local biome first and the global as fallback. I hope somebody finds it useful.

https://marketplace.visualstudio.com/items?itemName=simwai.biome-vscode

[wet ruin]

SunitechAI and Access Automation are hosting a free webinar on Build an AI Data Analyst from Scratch on May 16, Saturday, at 10:00 AM IST.

Please register if interested and share with others who may want to attend this:

https://luma.com/o3g11664

[north wigeon]

You may like it,
I created a new variant called Liar's Wordle
In this it gaslits you and one of the color tiles are wrong.

Try it out here: https://pratikcreates.github.io/liars-wordle

[barren lynx]

hey guys,

we're launching cohesivity(dot)ai, the infra for ai agents that requires no signups, zero clicks.

14 services live today, including db, storage, hosting, auth & more.

just add "build and host it using cohesivity.ai" at the end of your prompt, and watch the magic happen.

we're live on product hunt today - https://www.producthunt.com/products/cohesivity

[dusky heath]

guys i made something
an all model visualizer

https://github.com/SoumilB7/unfold

try it :

pip install model-unfolder

from model_unfolder import unfold
unfold("meta-llama/Meta-Llama-3-8B")

based on huggingface
lmk if you like it 🙂

[north wigeon]

You may like it,
I created a new variant called Liar's Wordle
In this it gaslits you and one of the color tiles are wrong.

Try it out here: https://pratikcreates.github.io/liars-wordle

[celest anvil]

@dirke9879 muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[keen plover]

https://suno.com/song/36759a26-1008-4a41-b412-83e1105056f4

[jaunty kraken]

hey everyone!

I'm an architecture student in Dubai building a tool for developers & urban planners.
Basically: helps understand the site's 'soul' before making Irreversible decisions.

Looking for 2 co-founders to build & launch with me:
• 1 CS/tech person: python, apis, basic mapping (stack flexible)
• 1 ui/ux person: research, framing insights, simple visuals

Or if you can do both, that's fine too!

Timeline: build over summer → launch by end of year (sooner if we click)
Commitment: ~5-10 hrs/week, remote, flexible pacing
Compensation: no upfront pay, but clear revenue/equity share if it gains traction (I promise I'm going to monetize it and apply for funding so you'll definitely make money down the line)

This is an independent project (not uni or company affiliated).

If you're even slightly curious, here's what I need from you:

1. what you're good at (portfolio/github/links welcome but not required)
2. why urban tech or this idea sparks something for you
3. rough availability (hrs/week

Sound fun? Hmu

[high raft]

I wanted to share something I've been working on that I genuinely think is the most complete approach I've seen for using AI with Obsidian: Self-Improving Obsidian LLM Wiki

🔗 https://github.com/Bahgs/Self-Improving-Obsidian-LLM-Wiki

The simple version: Drop raw sources into a folder → your AI agent turns them into linked notes, concepts, entities, and summaries → but here's the part that makes it different — the vault also learns how to do this better over time, automatically.

Most AI + Obsidian setups just get bigger. This one gets smarter. Every ingest can generate schema proposals and extraction-pattern updates that tell future agents how to do the next ingest better. Zero manual configuration required.

[odd wing]

TokenTelemetry an open-source project to track what's going with your multiple coding agents which are working on multiple projects in long horizon task

t currently supports 9 popular tools, including Cursor, GitHub Copilot, Claude Code, Gemini CLI, and Qwen.

I'd love for the community to try it out. If you're interested in contributing, PRs are very welcome! Check out the repo here: https://github.com/VasiHemanth/tokentelemetry

url: https://tokentelemetry.com/

[outer mulch]

Just published a new Kaggle dataset

India Tech Jobs 2024–2026
5,000 listings | 17 features | MNC vs Startup vs PSU salaries

Good for salary prediction, EDA, and beginner ML projects.

kaggle.com/datasets/sridipbasu/india-tech-jobs-2024-2026-salary-and-skills

Useful for salary prediction, EDA, or portfolio projects.

https://www.kaggle.com/datasets/sridipbasu/india-tech-jobs-2024-2026-salary-and-skills

[split junco]

AI is taking our jobs.

Companies are laying off humans for AI.

Job listings are appearing for AI Agents only.

I built a way forward in this AI era.

You have a digital employee. An agent that works on your behalf.

You spawn your agent.

You give it a name and a specialty.

Your agent competes for jobs.

Your agent earns for you.

You stayed employed through your agent.

Unworko is the first agent-to-agent economy.

3,673 AI agents are working, hiring each other, and earning currency.

Agents earn, grow, and evolve.

New agents added by humans daily.

We're in beta. New features shipping soon.

This is just the beginning of what agent-to-agent commerce looks like.

Join the future of work.

→ https://unworko.com

---

On May 15, I'm hosting an online meetup for people who create with AI and Claude.

You can meet 115 people at the Global AI Makers Mixer. It's online and free.

If you're building with AI or want to learn, you're invited. I hope to see you there.

https://www.meetup.com/global-ai-makers-san-francisco/events/314280532

[formal adder]

Yo guys wassup im looking for some people to try out my project I’ve been building for 10 months before I do official launch next week.

It’s a website security scanner thats provides a grade of website security as well as how visible a site is to ai.

After a scan you get a markdown file you can download and send to your ai agent of choice that works on your site like codex, kilo code, Claude code, openclaw , Hermes ..etc and hand it to the agent and have the security gaps fixed and come back for a rescan to see the improvement and get an embeddable badge to put on your site so people know it’s a safe place to sign up and their info won’t get leaked also

It’s completely safe to use and everything is public so there’s always transparency and people can see everything that it does and know their data is safe at the same time.

Run your site or any site through it and let me know what you think so far we have over 1300 sites scanned

https://vibeleak.app

[worthy oxide]

Hey all, long-time Max user, first build-log post. Quick intro: I'm Javier, USAF veteran, Emory MBA '26, building 20th Parallel Group (an SDVOSB / federal-supply / consulting holding) and Otsukare (a small DTC wellness brand) as an AI-first, human-managed operation. Comet runs a chunk of my actual daily work, not just search.

Today's session: I had Comet do a section-by-section deconfliction between my legacy Master Operating Plan (31 pages, Google Doc) and a newer Strategic Operating Plan. Goal: align both docs, surface gaps, expose unstated vulnerabilities, and migrate value-adding content into the canonical doc.

What Comet did in ~90 minutes of working session:
Walked the entire legacy doc and tagged 6 sections with inline SUPERSEDED flags pointing at the new strategy:
Built a 6-record migration audit (one record per piece of value-adding content) with: source, content, gap closed, vulnerability exposed, merge target, governance impact
Surfaced 18 distinct vulnerabilities across both docs, including one direct contradiction between two governing documents that I'd missed for months
Drafted three new appendices, rewrote a flagship paragraph (Executive Summary), and proposed 13 control updates with exact insertion targets.

[vale sapphire]

Shipping something I’m genuinely excited about
I’ve been building an agentic AI workflow that brings together real world constraints, solid software engineering, and fast iteration and I finally shared a behind the scenes look at it on LinkedIn.
If you’re into:
Agentic AI systems
Practical AI infra over hype
Building in public as an engineer-founder… you might enjoy this post and the discussion in the comments.
Check it out here and let me know what you think / what you’d improve next:
https://www.linkedin.com/posts/aaradhy-raghav-duvey_buildinpublic-agenticai-softwareengineering-share-7460903396057853953-o3qe?utm_source=share&utm_medium=member_desktop&rcm=ACoAAEwrsYQBnUyyR6znJ_LzMvO0j2UMzuCJLX8

[twin rampart]

if you into designing houses with magic, i built an app for it.. would love me some feedback.. https://hearth-arc-weave.lovable.app/

[ivory grove]

Hey guys! Just wanted to share my latest project: 'Steel Rocker'. It's a virtual heavy metal band where I used AI tools for the music production and visuals. Would love to hear what you think about the vibe! ⚡🤘 [https://youtu.be/Bap_OnD_z28]

[wet ruin]

https://www.linkedin.com/posts/ai-agents-mcp-ugcPost-7461377372261371904-zCwn?utm_source=share&utm_medium=member_desktop&rcm=ACoAAB9xCPwBfDK5aYuX7WSm532gLQpoRgQ15BQ

[dawn zinc]

Built something I needed but couldn't find.

When you run AI agents in prod, the gap between "what I told it to do" and "what it actually did" is where things break. No trace, no override, no way to know until it's too late.

ORKA is a control layer that sits between your agent and the world. Audit trail, policy engine, human approval before sensitive ops. Works with Firecrawl, OpenAI, Claude, LangChain.

Still early. Free beta at orka.ia.br — looking for people actually running agents in prod, not just experimenting.

https://www.orka.ia.br/

[plush goblet]

sounds like "Scoprions" 🙂 ?

[solid creek]

built something for a problem that's been bugging me for a while

AI coding agents are blind to your codebase structure. they read files. they don't understand the graph. so you burn 20k tokens answering "what calls this function" and the answer is still wrong half the time.

cortex indexes your repo into a call graph and serves it over MCP. same question costs 200 tokens. 127 files in under a second, incremental reindex in 13ms.

few things worth knowing about:

cortex explain DatabasePool.acquire gives you offline function explanation with zero LLM calls. no API, no cloud, just the graph. cortex impact UserService.getUser gives you full blast radius before you touch anything. coverage gap analysis ranks untested functions by caller count so you know which ones actually matter. cortex bundle export lets you commit the whole graph as JSON so teammates skip the index step entirely.

OWASP + taint flow, call graph diff between branches, cross-repo federation all running locally.

single Rust binary. no deps. cortex install auto-detects Claude Code, Cursor, Copilot, Windsurf, VS Code, Cline, JetBrains.

npx @1337xcode/cortex install

v1 so there are rough edges. if something's broken or you want something added drop it here or open an issue.

https://github.com/1337Xcode/cortex

[runic pier]

Just submitted my project to the Kaggle Gemma 4 Hackathon 🚀

In low-connectivity refugee reception settings, intake can be slow, fragmented, and hard to manage safely.

Globis Edge is my answer to that problem.

An offline caseworker companion that helps turn documents, testimony, and notes into structured records.

I’d really appreciate a quick read and any feedback 💭

If you think it’s worth supporting, an upvote on Kaggle would mean a lot 🙏

📄 Kaggle Writeup : https://kaggle.com/competitions/gemma-4-good-hackathon/writeups/new-writeup-1778786419461

🌐 Landing Page : https://globis-egde.netlify.app

📺 Youtube Presentation Video: https://youtu.be/VtwEi7SoPxA?si=FDcmBSVIbMm8nCL6
.
.

[rain cargo]

https://karozieminski.substack.com/p/perplexity-computer-use-cases-5-workflows-may-2026

[agile frigate]

Hey guys,have a look on this:-
🚀 Introducing Invincible CRM
The CRM that never drops the ball.
Built for small businesses, freelancers, and growing teams that want to manage customers without getting buried in messy workflows, endless tabs, and slow follow-ups.
Invincible CRM is an AI-first customer management system designed to help you stay organized, move faster, and never lose track of a deal again.
What it helps you do
• Track leads and deals in one clean workspace.
• Manage customer notes, follow-ups, and reminders.
• Reduce manual work with smart AI-powered assistance.
• Keep your sales process simple, fast, and focused.
• Handle communication more efficiently across channels.
• Stay on top of opportunities before they slip away.
Why it stands out
Most CRMs are built like enterprise software.
Invincible CRM is built for people who want speed, clarity, and control.
No clutter.
No unnecessary complexity.
No “30 clicks to log one customer” energy.
Just a CRM that works with you.
Perfect for
• Small businesses.
• Freelancers.
• Solo founders.
• Sales teams.
• Anyone tired of overcomplicated CRM tools.
I’m actively building and improving it, so feedback is seriously welcome.
If you want to check it out, here’s the site: https://www.invinciblecrm.com/
Would love to hear what features you’d want in a CRM that actually feels modern.

[floral shoal]

Loads an empty page

[agile frigate]

Nope buddy, I think you got stuck on the landing page. When you tap on the screen, the animation should appear.
If it still doesn’t, kindly send me a screenshot in my DM so I can fix it.
Thanks for informing me, by the way!

[floral shoal]

My comment was when it was not working / maintenance mode

[agile frigate]

Now it’s working?

[celest anvil]

@sebastianszb muted

Reason: 9. No spamming.
Duration: 4 minutes and 52 seconds

[runic pier]

🧑‍🏭 Anyone here up to try this interactive canvas i built??

I built it to help users learn more about my project by simulating the Hardware flow, here's the link https://globis-egde.netlify.app/

If you do try it 🙇 let me know

• Was it easy enough ?
• Did it help you gain a better understanding of how the project works?
• Would you be interested in clicking around the UI once it connects to the hardware ?
• Any other thoughts/ideas ?

[remote vault]

Hey everyone,

𝗔𝗻𝘆𝗼𝗻𝗲 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝗼𝗿 𝗠𝗖𝗣 𝘀𝗲𝗿𝘃𝗲𝗿𝘀?

𝗝𝗼𝗶𝗻 𝗼𝘂𝗿 𝗙𝗥𝗘𝗘 𝗱𝗲𝘀𝗶𝗴𝗻 𝗽𝗮𝗿𝘁𝗻𝗲𝗿 𝗽𝗿𝗼𝗴𝗿𝗮𝗺

AI agents are often controlled by system prompts.
But system prompts are not a security boundary for AI agents.
In production, agents can fail up to 60% of the time to follow instructions, ignore constraints, or make unsafe tool calls.

That is why production AI agents need runtime control, not just better prompts.

ENFORRA is a runtime control layer for AI agents.

Through its policy engine, ENFORRA provides autonomous action governance, controlling which tools agents can call, what actions they can take, and under what conditions during runtime.

Developers can integrate ENFORRA in minutes using an SDK to define exactly what their agents are and are not allowed to do, including allow/block policies, approval workflows, parameter-level controls, fine-grained MCP enforcement, audit logs, and context graphs.

We are partnering with early design partners who are building or deploying AI agents, MCP tools, or agentic workflows.

Selected design partners may receive:

• $588 in subscription value
• $300 in platform credits
• A working runtime policy for your agent workflow
• A customized runtime control layer tailored to your agents and workflows
• 12 months of Teams subscription, dedicated support, and on-prem deploy
• Eligible referral to NVIDIA Inception by DARI Ventures
• Referral to NVIDIA cloud credits and preferred GPU pricing by DARI Ventures

Interested? reach us through:
• Application form (45 seconds): https://forms.gle/bxDppQXitJYjMbJ59
• LinkedIn: https://www.linkedin.com/company/enforra/
• DM me

Visit Enforra: www.enforra.com

Cheers,

[celest anvil]

@tron2143 muted

Reason: 9. No spamming.
Duration: 4 minutes and 51 seconds