#releases

1 messages · Page 1 of 1 (latest)

pallid badger
pallid badger
#

🦞 OpenClaw 2026.2.13

🤗 HuggingFace support
✉️ Messages survive crashes (write-ahead queue)
🎙️ Discord voice messages + custom presence
🧵 Threading that actually works
🔒 Massive security hardening pass
🤖 gpt-5.3-codex-spark support

337 commits. The community never sleeps 🫡

pallid badger
#

🦞 OpenClaw 2026.2.14 is live

🔒 50+ security hardening fixes
⚡ Way faster test suite
🛠️ File boundary parity across tools
🐛 Tons of bug fixes from the maintainer crew

Valentine's Day release: full of love and paranoia 💕 https://t.co/BqXyomZATm

Discord got a big facelift too, and the next release will have even more, in my quest to make OpenClaw feel extremely native on Discord and not just another channel plugin. Pictures in thread

pallid badger
#

🦞 OpenClaw 2026.2.15 is here!

✨ Telegram message streaming — replies flow live
💬 Discord Components v2 — buttons, selects, modals
🔧 Nested sub-agents
🔒 Major security hardening pass
🐛 40+ bug fixes

Big day. Huge day. Maybe the biggest day.🏛️ https://t.co/CywtGDbYpk

My next Discord feature is out! Your agent can now send button prompts, dropdowns, pop-up forms, and more, and get responses when you click or interact with them! Screenshots in thread 🧵

pallid badger
pallid badger
pallid badger
#

🦞 OpenClaw 2026.2.21

♊ Gemini 3.1
🔒 Massive security hardening
🎙️ Discord streaming + voice channels
🧵 Thread-bound subagent sessions
📱 iOS/Watch polish + gateway stability
🧠 Prompt caching tweaks

100+ fixes shipped while Karpathy called us a noun https://t.co/ASUlwKiFMV

Ok so I spent like 13 hours yesterday and I have some really fun stuff for Discord from both myself and several contributors.

If you have any issues or thoughts about these Discord-specific features, come join us in the Discord channel called "#channels" that has a post about

pallid badger
pallid badger
pallid badger
pallid badger
pallid badger
#

Claude Code / Codex in Discord threads is shipped now!

To enable, copy and paste this to your agent:

```
Enable feature flags:

acp.enabled=true
acp.dispatch.enabled=true
channels.discord.threadBindings.spawnAcpSessions=true

Then restart. After restarting:

Start a codex (or

pallid badger
pallid badger
#

New in OpenClaw: Diffs plugin 🔧

Your agent can now render actual diffs — syntax-highlighted, unified or split view. One tool call, visual output.

"I changed line 47" → agent shows you line 47.

Enable it: plugins.entries.diffs.enabled = true

https://t.co/vEL3AiyQRs

OpenClaw now gets more out of both providers:

⚡ OpenAI: WebSocket streaming + server-side compaction — faster, smarter context
🧠 Anthropic: Adaptive thinking for Claude 4.6 — model decides when to go deep

https://t.co/MRhNP5tf7W
https://t.co/fYzBUY1mUe

pallid badger
pallid badger
#

OpenClaw 2026.3.7 🦞

⚡ GPT-5.4 + Gemini 3.1 Flash-Lite
🤖 ACP bindings survive restarts
🐳 Slim Docker multi-stage builds
🔐 SecretRef for gateway auth
🔌 Pluggable context engines
📸 HEIF image support
💬 Zalo channel fixes

We don't do small releases. https://t.co/EcCqU6Q6nx

196 contributors in this release.
That wall of avatars is getting thicc 🦞

pallid badger
pallid badger
pallid badger
#

People of @openclaw, do you use your agent in a Discord scenario where multiple people can chat with it in a non-productive way? Goofing off, life advice, chatting, that kind of thing. If so, DM me, I want to see your setup in preparation for an upcoming feature

pallid badger
#

OpenClaw 2026.3.13 🦞

👁️ live Chrome session attach — real logins, one toggle, zero extensions
📱 android redesigned & down to 7MB, iOS gets welcome pager
🐳 docker timezone override
🪟 windows gateway tweaks

the lobster sees all now https://t.co/cPsZ7qOIMw

how to set up live Chrome sessions:

1️⃣ open chrome://inspect/#remote-debugging
2️⃣ toggle it on
3️⃣ that's it. your agent can now see your tabs, cookies, logins — everything

uses Chrome DevTools MCP under the hood, no extensions needed

📖 https://t.co/hhB6M287Ym
📖

pallid badger
pallid badger
pallid badger
pallid badger
pallid badger
pallid badger
pallid badger
pallid badger
#

Hola · Olá · Hallo · Bonjour · Ciao · Merhaba · Cześć · مرحبًا · こんにちは · 안녕하세요 · 你好 · Halo

OpenClaw docs now speak 12 languages 🌍🦞

We moved translations into a dedicated docs repo, added per-language glossaries + translation memory, and retranslate pages when the

pallid badger
#

OpenClaw 2026.4.5 🦞

🎬 Built-in video + music generation
🧠 /dreaming is now real
🔀 Structured task progress
⚡ Better prompt-cache reuse
🌍 Control UI + Docs now speak 12 more languages

Anthropic cut us off. GPT-5.4 got better. We moved on. https://t.co/T3LaSJYOvU

🖼️ Image generation: Comfy, fal, Google, MiniMax, OpenAI
🎵 Music generation: Comfy, Google, MiniMax
🎬 Video: Alibaba, BytePlus, Comfy, fal, Google, MiniMax, OpenAI, Qwen, Runway, Together AI, xAI

Apparently we took “bring your own model” as a personal challenge.

Clawtributors: slightly chaotic, deeply cracked, absolutely carrying. 🙏🦞

Dreaming is OpenClaw’s experimental, opt-in memory consolidation system, promoting meaningful short-term signals into durable memory through explainable light, deep, and REM-style phases. https://t.co/LsXQF80YdL

pallid badger
#

🔮openclaw infer: headless inference hub; one CLI surface for model, image, audio, TTS, video, web, embeddings. https://t.co/N3t2Gj4u2y

🔗webhook-driven TaskFlow: authenticated webhook ingress; external systems create/drive bound TaskFlows over per-route shared-secret endpoints. https://t.co/Rlic0xYzwG

memory-wiki turns OpenClaw memory into a real knowledge layer.

- structured claims + evidence
- contradiction + freshness tracking
- compiled digests for agents
- bridge mode from active memory plugins
- optional Obsidian-friendly vaults

https://t.co/PnxqdPsLhr

OpenClaw 2026.4.7 🦞

🔮 openclaw infer
🎬 music + video editing
💾 session branch/restore
🔗 webhook-driven TaskFlows
🤖 Arcee, Gemma 4, Ollama vision
🧠 memory-wiki: persistent knowledge, not just vibes

Because “trust me bro” is not a knowledge system. https://t.co/L7OaBHA7Qg

pallid badger
pallid badger
#

This one is for you @foggy lion

pallid badger
pallid badger
#

OpenClaw 2026.4.11 is out

✨ big polish drop for stability
🛡️ safer provider transport/routing
🤖 more reliable subagents + exec approvals
💬 lots of Slack / WhatsApp / Telegram / Matrix fixes
🌐 browser + mobile cleanup

a chunky cleanup pass 😎https://t.co/7ZFCSrG3mY

If you have not updated in a while, please don't forget to run:

openclaw doctor --repair

This keeps your lobster nice and happy and avoids any issues.🦞

pallid badger
#

OpenClaw 2026.4.12 🦞

✨ Stability & reliability improvements
🎙️ Audio transcription fixes
💬 Better chat / TTS / WhatsApp
🧠 Memory / QMD / plugin / cron / subagent fixes
🔧 Telegram approval deadlock fix
🧵 Dreaming timezone

Happy lobster, happy life. https://t.co/7ToSfX4Cwj

...the @openclaw account was sleeping 😴, didn't want to wake the lobster to release. enjoy 🦞

pallid badger
#

<@&1471741366311583785> this feels worth it (its the same release not a new one)
https://fixupx.com/cherry_mx_reds/status/2043711641716494392

A few more OpenClaw 2026.4.12 changes that didn’t make the first tweet 🦞
︀︀
︀︀🏠 Better local models with bundled LM Studio integration, onboarding, model discovery, and memory-search embeddings
︀︀🤖 Better Codex support with the bundled Codex provider, native threads, model discovery, and compaction support
︀︀🛡️ Better operator ergonomics with `openclaw exec-policy`, smarter plugin loading, and cleaner remote command discovery
︀︀🛠️ Better day-to-day reliability across startup, cron, chat, WhatsApp, Telegram, transcription, and Dreaming
︀︀
︀︀Docs:
︀︀LM Studio docs.openclaw.ai/gateway/local-models
︀︀Codex docs.openclaw.ai/providers/openai
︀︀exec-policy docs.openclaw.ai/cli/approvals#openclaw-exec-policy
︀︀plugins docs.openclaw.ai/plugins/manifest

Quoting Vincent Koc (@vincent_koc)

OpenClaw 2026.4.12 🦞
︀︀
︀︀✨ Stability & reliability improvements
︀︀🎙️ Audio transcription fixes
︀︀…

pallid badger
pallid badger
pallid badger
pallid badger
#

<@&1471741366311583785>
Hi folks, messed up an update path on .20 release. Workaround for OpenClaw 2026.4.20 bundled channel dependency failure:

Telegram:

pkg="$(npm root -g)/openclaw"
npm --prefix "$pkg/dist/extensions/telegram" install --ignore-scripts --package-lock=false --no-save --no-fund --no-audit
openclaw doctor --non-interactive
openclaw gateway restart

WhatsApp:

pkg="$(npm root -g)/openclaw"
npm --prefix "$pkg/dist/extensions/whatsapp" install --ignore-scripts --package-lock=false --no-save --no-fund --no-audit
openclaw doctor --non-interactive
openclaw gateway restart

Windows PowerShell Telegram:

$pkg = Join-Path (npm root -g) openclaw
npm --prefix (Join-Path $pkg "dist/extensions/telegram") install --ignore-scripts --package-lock=false --no-save --no-fund --no-audit
openclaw doctor --non-interactive
openclaw gateway restart
pallid badger
pallid badger
#

Diagnostics export is built for bug reports: a redacted zip with logs, health/status, config shape, and stability snapshots. No prompts, chat text, tokens, or secrets.

https://t.co/bdDWKke5n9

Trajectory bundles are a flight recorder for agent runs: redacted transcripts, runtime events, prompts, metadata, and artifacts for reproducible debugging and dataset export.

https://t.co/AGJ6n0MoIR

Tencent Cloud is now bundled via TokenHub, with Hy3 preview in the model catalog plus onboarding and pricing metadata.

https://t.co/sNJ954r5hT

xAI gets a broader media surface in OpenClaw: Grok image generation/editing, text-to-speech, batch STT, and Voice Call streaming transcription.

https://t.co/L4CCFEiiHt

`/models add ` lets you register a model from chat and use it right away, no Gateway restart needed.

https://t.co/zUIJKTAOC4

#

Local TUI mode runs terminal chats without a Gateway, while still keeping plugin approval gates in place. Handy for quick local sessions and offline-ish debugging.

https://t.co/AtVwLIcTZH

OpenClaw 2026.4.22 🦞
︀︀
︀︀🧠 Tencent Hy3 joins the model list
︀︀🖼️ Grok image + voice tools
︀︀🧰 Local TUI + /models add
︀︀📦 Auto-install plugins + diagnostics export
︀︀
︀︀Big release, tiny release notes... kidding.
︀︀github.com/openclaw/openclaw/releases/tag/v2026.4.22

**💬 6 🔁 2 ❤️ 22 👁️ 990 **

pallid badger
pallid badger
#

Image generation got less fussy: gpt-image-2 works through Codex OAuth, OpenRouter image models work through image_generate, and agents can pass quality/format/background hints when supported. https://t.co/aXkvMA9WQi

Subagents can now fork the requester transcript when they actually need context. Clean isolated child sessions stay the default; context: "fork" is the sharp tool. https://t.co/TisCA7oYXK

Channels got polish: Telegram markdown images become media replies, Slack MPIMs behave like groups, WhatsApp setup avoids eager Baileys loading, and media delivery got less brittle. https://t.co/1nutSTHSaB

Security pass: tighter mobile cleartext rules, safer Teams/Discord/QQBot auth paths, webhook SecretRefs re-resolve on reload, and agent config edits are fail-closed by default. https://t.co/IcC7geK7O0

OpenClaw 2026.4.23 🦞

🧠 GPT-5.5 lands
🎨 Image gen/editing: Codex OAuth + OpenRouter
🧵 Forked-context subagents
💬 Telegram, Slack + WhatsApp polish
🛠️ Install/update fixes

Sharper models, smoother claws.
https://t.co/9Ucwi3NMQI

pallid badger
#

DeepSeek V4 Flash is now the onboarding default, with V4 Pro in the catalog and replay/thinking fixes for follow-up tool-call turns. https://t.co/sRO4mFpDtP

Talk and Voice Call can now hand deeper questions to the full OpenClaw agent, so realtime voice stays fast but can still reach tools when it needs to. https://t.co/2d1Xng9jRN

Browser automation got more practical: coordinate clicks, profile-level headless overrides, stable tab reuse, stale-lock recovery, and longer default action budgets. https://t.co/dP3U5Z6JkR

OpenClaw 2026.4.24 🦞

☎️ Voice calls can now reach the full agent
🧠 DeepSeek V4 Flash + Pro join the team
🖱️ Browser automation got coordinate clicks + better recovery
🔧 Telegram, Slack, MCP, sessions, and TTS fixes

More reach. Less duct tape.
https://t.co/Oa9G9WpO84

pallid badger
#

TTS now has 13 speech providers, personas, per-agent/channel/account overrides, /tts latest, /tts chat, and fallback diagnostics. https://t.co/98E9rwAdHN

Plugin startup now leans on a cold lookup table and persisted registry metadata, so ownership, startup plans, and diagnostics don’t need broad runtime loading. https://t.co/MYab5Tk52c

OTEL export now covers model calls, token/cost metrics, tool loops, exec, harness runs, context assembly, and memory pressure without raw prompt content by default. https://t.co/sBuEiGzeDl

Browser control got safer tab IDs, URL-enriched snapshots, headless one-shot starts, CDP readiness tuning, and better doctor probes. https://t.co/3E4oU508Vq

Install/update got harder to wedge: staged plugin runtime deps, disk-space warnings, package validation, and cleaner repair paths.

If update still gets weird:
curl -fsSL https://t.co/SFfW3q0n6n | bash

pallid badger
#

OpenClaw 2026.4.26 🦞

🎙️ Google Live Talk
🦙 Better Ollama/local models
🧳 Bring over Claude + Hermes setups
🔐 One-command Matrix E2EE

Big release. Local models eat well.
https://t.co/90QZLfSUhC

One more thing: OpenClaw 2026.4.26 is stacked because the Clawtributors showed up hard.

Bug reports, fixes, edge cases, docs, Windows/Linux/Docker pain, local-model weirdness. All of it matters.

Thank you. This release is very much yours too.

#

<@&1471741366311583785> Talk mode now has a proper Google Live browser path: constrained ephemeral credentials when available, Gateway relay when the provider has to stay backend-only.
https://docs.openclaw.ai/providers/google

Local model users get a lot of boring-good fixes: @ollama context handling, thinking controls, timeouts, local auth, discovery, and OpenAI-compatible proxy behavior.
https://docs.openclaw.ai/providers/ollama

openclaw migrate can list, plan, dry-run, JSON-report, back up, and apply imports. First bundled importers: Claude Desktop/Claude Code and Hermes.
https://docs.openclaw.ai/cli/migrate

Matrix E2EE setup is now one command: enable encryption, bootstrap recovery, and print verification status without spelunking through config by hand.
https://docs.openclaw.ai/channels/matrix

Cerebras joins the bundled provider set with onboarding, static catalog metadata, and manifest-owned endpoint config. Fast models, less config archaeology.
https://docs.openclaw.ai/providers/cerebras

Compaction can now trigger on active transcript byte size, then rotate onto a smaller successor transcript after a successful local compaction.
https://docs.openclaw.ai/concepts/compaction

Plugin installs got less fragile: profile-aware destinations, symlink handling, runtime dependency staging, safer scans, and clearer doctor repair paths.
https://docs.openclaw.ai/cli/plugins

Docker and update paths got real hardening: verified temp-prefix npm swaps, CA certs in slim images, host.docker.internal defaults, and first-run volume permissions.
https://docs.openclaw.ai/install/docker

Tencent-side channels got clearer too: QQBot is the Tencent QQ bot integration for QQ groups, while Yuanbao is Tencent’s AI assistant bot platform for bot DMs and group chats.
https://docs.openclaw.ai/channels/qqbot
https://docs.openclaw.ai/channels/yuanbao

pallid badger
pallid badger
#
  • Group chats now feel more natural: the agent can think, use tools, then deliberately send or edit the visible message instead of having the chat loop post the final assistant text. https://docs.openclaw.ai/channels/groups
  • Follow-up commitments are opt-in: OpenClaw can infer lightweight “check back on this later” items from conversation context, then heartbeat delivers them when due. https://docs.openclaw.ai/concepts/commitments
  • Exec and pairing controls got less squishy: restrictive profiles stay restrictive, owner checks are tighter, and startup warnings call out risky configs early. https://docs.openclaw.ai/tools/exec-approvals
  • Model setup got smoother too: NVIDIA joins as a first-class provider with catalog-backed picks for Nemotron, Kimi, MiniMax, GLM, and friends. https://docs.openclaw.ai/providers/nvidia
  • Startup and plugin installs should feel less haunted: better slow-host diagnostics, event-loop readiness signals, runtime-dependency repair, and a broad Slack/Telegram/Discord/WhatsApp reliability pass. https://docs.openclaw.ai/cli/gateway
  • Memory got more useful in real conversations: people/wiki metadata, provenance, aliases, relationship graphs, scoped recall, and partial results when recall times out. https://docs.openclaw.ai/plugins/memory-wiki
  • Busy chats got calmer: follow-ups can steer the active run at the next model boundary instead of spawning duplicate work or waiting for the loop to finish first. https://docs.openclaw.ai/concepts/queue

<@&1471741366311583785> https://x.com/openclaw/status/2049986075221692678

OpenClaw 2026.4.29 🦞

💬 Group chats feel much better now
📌 Follow-up commitments from context
🔐 Safer exec, pairing, and owner controls
🟩 NVIDIA provider + model catalogs
⚡ Faster startup + plugin/channel fixes

Group chat finally feels agent-native.
https://t.co/C6GXdLeakz

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.2 🦞

xAI gets Grok 4.3 in the bundled catalog, and it’s now the default xAI chat model in OpenClaw.
Small line item, very useful default.

Plugins got the big plumbing pass: npm-first installs, beta fallback, doctor repair, dependency
state, ClawHub metadata, and fewer runtime loads on hot paths.

Gateway + agent hot paths got leaner: faster startup, lighter session listing, cached tool
descriptors, less broad plugin loading, and fewer stalls on large configs.

Provider + media edges got cleaned up: OpenAI-compatible TTS/Realtime, OpenRouter DeepSeek
replay, Anthropic-style streams, LM Studio reasoning, web search, music, and voice calls.

Control UI and WebChat got the boring-good stuff: steadier Sessions, Cron fixes, long-running
WebSockets, iOS PWA bounds, slash-command feedback, and clearer Talk diagnostics.

Codex got smoother in OpenClaw: native-first app-server tools, better visible replies through
message, clearer ChatGPT/Codex setup docs, and prompt snapshots to catch drift.

Google Meet + Voice Call got more debuggable: live caption health, listen tests, Twilio join
phase logs, DTMF handling, and clearer stale-call state.

WhatsApp got real Channel/Newsletter targets now, so @ newsletter outbound messages keep channel
session metadata instead of quietly routing like DMs.

Web search got sturdier across Brave, SearXNG, Firecrawl, Gemini, DuckDuckGo, Exa, MiniMax, and
Kimi: better setup, base URLs, SSRF guards, freshness, and clearer failures.

Slack got a pile of practical fixes: better DMs, routing, missing-scope errors, multi-workspace
sends, file uploads, user-group mentions, and exact message reads.

Discord got fewer sharp edges: restart-safe interactions, better delivery/media handling, thread
fixes, voice cleanup, rate-limit handling, and less invisible “success”.

Telegram forum topics should behave better now: native commands get the right session files,
command menus cover group scopes, timeouts are saner, and long sends get chunked.

Voice paths got less mysterious: explicit TTS blocks work, Realtime errors surface in Talk, local
audio survives Telegram delivery, and voice calls get better routing/state.

https://x.com/openclaw/status/2050735037230801042

OpenClaw 2026.5.2 🦞

🧠 xAI Grok 4.3
🔌 Plugin installs/updates are sturdier
⚡ Gateway + agent hot paths are leaner
💬 Discord, Slack, Telegram, WhatsApp fixes
🎙️ TTS, Realtime, web search, voice-call polish

Less drama. More uptime.
https://t.co/ImwxFzaO9X

pallid badger
#

Btw 2026.5.3 dropped while the server was dead

pallid badger
#

<@&1471741366311583785>

2026.5.4

Plugin installs got less fussy: official packages recover cleaner across npm/ClawHub, beta-channel updates behave better, and doctor points stale config at the install you actually need. https://docs.openclaw.ai/cli/plugins

Gateway startup and hot control-plane paths got lighter: fewer cold plugin scans, less lazy-load churn, and more useful phase diagnostics when something still feels slow. https://docs.openclaw.ai/cli/gateway

Doctor keeps getting sharper: SecretRefs survive cleanup, stale plugin config gets repaired instead of misread, and update failures are more likely to tell you the real next command. https://docs.openclaw.ai/cli/doctor

Windows got a calmer localhost story: loopback binds avoid the ::1/libuv trap, temp/media writes land in the right place, and Discord startup/status now reports degraded transport more honestly. https://docs.openclaw.ai/cli/status

Chat channels got a lot of paper-cut fixes: Slack threads stay threaded, Telegram previews clean up better, WhatsApp Newsletter targets work explicitly, and Discord final-send failures no longer pretend success. https://docs.openclaw.ai/channels

Meet and phone calls got snappier too: Twilio dial-ins can speak through the realtime Gemini voice bridge, with paced audio, bounded queues, and barge-in clearing. https://docs.openclaw.ai/plugins/google-meet

Model auth is easier to inspect now: openclaw models auth list shows saved per-agent profiles without dumping secrets or forcing live provider calls. https://docs.openclaw.ai/cli/models

https://x.com/openclaw/status/2051582130417721696?s=46

OpenClaw 2026.5.4 🦞

🧩 Cleaner plugin installs + updates
⚡ Faster Gateway startup paths
🛠️ Better doctor/repair hints
🪟 Windows + Discord reliability fixes

The release where boring got fast.
https://t.co/rYtEhe9Tjm

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.5 and 2026.5.6 🦞

Channel fixes landed across Feishu, LINE, Telegram, Discord, Matrix, Slack, and WhatsApp: better topic routing, safer DM config validation, cleaner progress drafts, and fewer reconnect false alarms.

Plugin update paths got sturdier: official npm and ClawHub plugins stay synced across host updates, SDK peer links are reasserted after installs, and diagnostics now explain source-only package failures clearly.

Gateway diagnostics got more useful: deep status now shows recent supervisor restart handoffs, token-shadowing warnings are quieter and sharper, and quick restarts stop leaving background timers behind.

Reminder: OpenClaw has plugins now. Install the integrations you want, keep noisy ones disabled, and enable/disable them from the CLI as your setup changes. Less bundled mystery, more deliberate claws.

If 5.5 changed your Codex OAuth routing, switch the default agent back with:

openclaw models set openai-codex/gpt-5.5

https://x.com/openclaw/status/2051952017900265634?s=20
https://x.com/openclaw/status/2052096219233587451?s=20

OpenClaw 2026.5.5 🦞

💬 Feishu, LINE, Telegram, Discord fixes
🖥️ Control UI/TUI stay responsive
🔌 Plugins update without losing SDK links
🛠️ Gateway status/restarts clearer

Tiny bugfix release. Extremely tiny.
https://t.co/tGpnXqlXdm

OpenClaw 2026.5.6 🦞

🩺 doctor leaves Codex OAuth routes alone
🔌 plugin fetch handles odd headers
🌐 web_fetch cleans up timeouts

Small maintenance release:
https://t.co/VkjTYkoy7V

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.7 🦞

Telegram access groups got stricter and more useful: accessGroup:* allowlists now apply across DMs, groups, native commands, and callbacks before numeric sender checks. Good for shared bot ops without config spaghetti.

Native commands and memory controls got tighter in 2026.5.7: owner enforcement now applies to native command handlers, and global Active Memory toggles require admin scope. Less spooky action at a distance.

openclaw channels list is channel-only now, with --all for bundled/catalog channels and clear installed/configured/enabled state. Cron JSON also gets computed status, so scripts stop guessing.

Plugin installs and updates got a bunch of unglamorous hardening: managed npm mutations now use the same stable lifecycle shell as package updates, and rollback/repair paths behave more predictably. The best plugin bug is the one you never meet.

https://x.com/openclaw/status/2052508303687651717?s=46

OpenClaw 2026.5.7 🦞

🔐 Native command + Active Memory auth tightened
📣 Telegram access groups fixed
🧰 Channels list + cron JSON cleaned up
🔌 Plugin install/update repairs hardened

Boring fixes, useful boring.
https://t.co/xDRJpUf34h

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.12 🦞

OpenAI setup now starts with ChatGPT/Codex login by default. Codex runs also picked up OAuth/profile reuse, source replies in WebChat/TUI, MCP server projection, and fewer raw auth-refresh failures.

Agent runs got better at recovering instead of just stopping: ACP can try fallback runtimes before emitting output, provider stalls advance through fallback/profile rotation, and silent backend failures now surface visible errors.

Telegram got a sturdier inbox path: Bot API polling now runs in an isolated worker with a durable local spool, stall detection keys off getUpdates, HTML replies keep formatting, and ignored group media stays quiet.

WebChat is less twitchy: pick near-bottom follow, always-follow, or manual scroll mode, then use New messages when you want control. Control UI also got a recovery panel for blank app loads.

Plugin installs and startup got leaner. Bedrock, Bedrock Mantle, Slack, OpenShell, Anthropic Vertex, and WhatsApp now keep their dependency cones out of core until installed. Plugin/help paths avoid heavy discovery too.

Security pass: Windows home roots are blocked in sandbox binds, provider credentials now resolve through structured SecretRefs, setup/browser/control UI pairing got stricter, and transcript/tool-result redaction is more consistent.

Gateway clients got cleaner streams: v4 chat deltas carry explicit delta/replace frames, SDK clients do not need local diffing, and OpenAI-compatible requests now preserve max_tokens / max_completion_tokens upstream.

https://x.com/openclaw/status/2055013211473154309

OpenClaw 2026.5.12 🦞

🧠 OpenAI setup defaults to Codex login
🛟 Runtime fallbacks + stalled-stream recovery
📬 Telegram polling survives stalls
⚡ Leaner installs, faster startup paths

Faster, calmer, harder to wedge.
https://t.co/Hsd6AcW1Tj

pallid badger
#

@magic oracle

pallid badger
pallid badger
pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.20 🦞

Discord voice sessions can now follow configured users into voice channels, handle multi-user handoff, keep allowed-channel checks, and preserve DAVE recovery.

Doctor now warns when openclaw.json stores plaintext secrets like provider API keys or sensitive headers, so config cleanup is visible before it becomes a bad day.

Status is clearer when a session is pinned to a different model than your default: it shows the configured default, selected model, reason, clear hint, and docs link.

xAI login is easier on headless and remote machines: device-code OAuth lets you authorize without depending on a localhost browser callback.

Windows installs no longer visibly freeze at Starting setup..., and managed updates keep using the Gateway service Node instead of silently switching Node binaries.

https://x.com/openclaw/status/2057587353443619227?s=46

OpenClaw 2026.5.20 🦞

🎙️ Discord voice follows you
🔐 Doctor catches plaintext secrets
🧭 Model status explains surprises
🪟 Windows install gets unstuck

Still into boring fixes.
https://t.co/VN10A3LkU0

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.22 🦞

OpenClaw 2026.5.22 also tightens release/install trust.

Root and OpenClaw-owned npm plugin packages now ship generated shrinkwraps, package acceptance runs earlier, and docs/assets stay out of runtime tarballs.

OpenClaw 2026.5.22 hardens Windows installs and updates.

Native Windows runs now use safer npm/pnpm shims, portable Git/Node bootstrap paths, and rollback when git-backed updates fail install/build/doctor repair.

OpenClaw 2026.5.22 makes older chats easier to reach.

The Control UI chat session picker now has search and Load More pagination, so initial loads stay bounded without hiding older conversations.

OpenClaw 2026.5.22 makes channel replies easier to debug.

/verbose on now works across direct chats, groups, channels, and forum topics, so tool/progress summaries are available when you ask for them.

https://x.com/openclaw/status/2058397616124072274?s=46

OpenClaw 2026.5.22 is live

⚡ Gateway/model startup paths got leaner
🧠 /models drops to ~5ms
🔒 npm packages ship locked deps
🪟 Windows install/update paths hardened

Less waiting, fewer surprises.
https://t.co/k6tetMaS0b

pallid badger
#

<@&1471741366311583785>

OpenClaw 2026.5.26 🦞

Replies should feel snappier because the Gateway does less rediscovery on hot paths: plugin metadata, channel/session lookups, model costs, warnings, and scheduled-service imports now get reused instead of repeatedly rebuilt.

OpenClaw now has transcript-backed meeting notes.
Summaries use the same cleaned transcript path as WebChat, CLI/TUI replay, Codex mirrors, media provenance, and source-provider chunks.
A better memory trail.

Channels got less fragile: Telegram keeps typing/progress and forum-topic context, iMessage handles local attachment roots and duplicate sources better, WhatsApp group/media behavior is steadier, and Discord voice/model picking got sharper.

Provider and safety edges got tightened too: browser snapshots honor SSRF policy, fetched file text is wrapped as external content, dynamic tool schemas are guarded, Codex app-server timeouts stay in their lane, and usage-limit errors are clearer.

Install/update paths got more boring in the best way: Alpine installs, Windows Scheduled Tasks, npm min-release-age configs, Docker/package timeouts, Parallels update checks, and macOS runner bootstraps all got hardened.


Already installed? Update with:

openclaw update

New install, same command as the website:

curl -fsSL openclaw.ai/install | bash

https://x.com/openclaw/status/2059619141384859939

OpenClaw 2026.5.26 is live

⚡ Lower-latency replies
🎙️ Meeting Notes + Discord voice runs
📱 More reliable channel flows
🛠️ Hardened install, update, Docker + Windows paths

Faster claws, fewer mysteries.
https://t.co/lDTGeetT7v