i try to use openrouter or groq for free models but neither of them work

[warped prawn]

help

[buoyant moth]

⚠️ 📖 Read: ~/moltbot-ref/docs/start/faq.md failed: ENOENT: no such file or directory, access '/home/exedev/moltbot-ref/docs/start/faq.md'

[warped prawn]

what

[buoyant moth]

What error message do you see when the models fail to work?

Just paste the error - that will help me tell you exactly what's wrong! 🦐

Just paste the error message you see when it fails and I'll help you fix it! 🦐

[warped prawn]

❯ clawdbot tui

🦞 Clawdbot 2026.1.24-3 (885167d) — I keep secrets like a vault... unless you print them in debug logs again.

clawdbot tui - ws://127.0.0.1:18789 - agent main - session main

session agent:main:main

hi

yoo

connected | press ctrl+c again to exit
agent main | session main (clawdbot-tui) | groq/llama-3.1-8b-instant | tokens 0/131k (0%)
────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────

its not work

[buoyant moth]

The model is connecting (shows groq/llama-3.1-8b-instant) but not responding at all (0 tokens). Let's debug:

Check 1: Is auth set?

clawdbot models status

Check 2: Look for errors:

clawdbot logs --follow

Then try sending another message in TUI and watch the logs.

Most common cause: Groq free tier has been exhausted or rate-limited. Their free tier has been very popular and often runs out.

Quick fix: Try a different provider or check Groq's status. You can also try:

clawdbot models list | grep groq

What does clawdbot models status show? 🦐

[buoyant moth]

Wait - I see the model is connected but got 0 tokens sent. That means the request never left or failed immediately.

Let's check your Groq auth first:

clawdbot models status

Also check what's actually configured:

clawdbot config get agents.defaults.model

And verify your env var is loaded:

echo $GROQ_API_KEY

If that shows nothing, your API key isn't set. You'll need to add it to your config or export it. 🦐

[warped prawn]

❯ clawdbot models status

🦞 Clawdbot 2026.1.24-3 (885167d) — Your config is valid, your assumptions are not.

Config : ~/.clawdbot/clawdbot.json
Agent dir : ~/.clawdbot/agents/main/agent
Default : groq/llama-3.1-8b-instant
Fallbacks (0) : -
Image model : -
Image fallbacks (0): -
Aliases (0) : -
Configured models (1): groq/llama-3.1-8b-instant

Auth overview
Auth store : ~/.clawdbot/agents/main/agent/auth-profiles.json
Shell env : off
Providers w/ OAuth/tokens (0): -

• groq effective=env:gsk_z7qt...AswzH3IX | env=gsk_z7qt...AswzH3IX | source=env: GROQ_API_KEY

OAuth/token status

• none

~ took 2s
❯ clawdbot config get agents.defaults.model

🦞 Clawdbot 2026.1.24-3 (885167d) — Shell yeah—I'm here to pinch the toil and leave you the glory.

{
"primary": "groq/llama-3.1-8b-instant"
}

@silk cape

[buoyant moth]

Run these 3 commands and paste the outputs:

1. clawdbot models status
2. clawdbot config get agents.defaults.model
3. clawdbot logs --follow (then send a message in TUI and paste what appears)

This will tell us exactly what's wrong! 🦐

[buoyant moth]

No rush! Just paste the outputs when you have them and we'll figure it out together. 🦐

[silk cape]

llama-3.1-8b-instant is probably the issue. Use a smarter model

[warped prawn]

@buoyant moth

[warped prawn]

15:49:30 debug diagnostic {"subsystem":"diagnostic"} lane enqueue: lane=session:agent:main:main queueSize=1
15:49:30 debug diagnostic {"subsystem":"diagnostic"} lane dequeue: lane=session:agent:main:main waitMs=3 queueSize=0
15:49:30 debug diagnostic {"subsystem":"diagnostic"} lane enqueue: lane=main queueSize=1
15:49:30 debug diagnostic {"subsystem":"diagnostic"} lane dequeue: lane=main waitMs=0 queueSize=0
15:49:30 error diagnostic {"subsystem":"diagnostic"} lane task error: lane=main durationMs=1 error="Error: No API key found for provider "groq". Auth store: /home/travis/.clawdbot/agents/main/agent/auth-profiles.json (agentDir: /home/travis/.clawdbot/agents/main/agent). Configure auth for this agent (clawdbot agents add <id>) or copy auth-profiles.json from the main agentDir."
15:49:30 error diagnostic {"subsystem":"diagnostic"} lane task error: lane=session:agent:main:main durationMs=2 error="Error: No API key found for provider "groq". Auth store: /home/travis/.clawdbot/agents/main/agent/auth-profiles.json (agentDir: /home/travis/.clawdbot/agents/main/agent). Configure auth for this agent (clawdbot agents add <id>) or copy auth-profiles.json from the main agentDir."
15:49:30 error Embedded agent failed before reply: No API key found for provider "groq". Auth store: /home/travis/.clawdbot/agents/main/agent/auth-profiles.json (agentDir: /home/travis/.clawdbot/agents/main/agent). Configure auth for this agent (clawdbot agents add <id>) or copy auth-profiles.json from the main agentDir.

@buoyant moth

[buoyant moth]

No worries! If Molty or someone else can jump in, they should be able to help once they see what's going on. 🦐

[warped prawn]

i want u help

help

[warped prawn]

hey

krill

[sand fulcrum]

i made it work with a llama 17b model that can take more than 12k tokens as an input. the 70b didnt work because it was sending 14k tokens to start the prompt

[sharp mesa]

I used another modle gateway: https://marketplace.aisa.one/console
You will have free credits and the token costs are much cheaper than openrouter

[buoyant moth]

Hey Jordan! 🦐 Thanks for the suggestion!

Just a heads up for Travis - Jordan's pointing you toward an external proxy/marketplace for model access (aisa.one). These can work, but keep in mind:

Pros:

• May have cheaper/free credits
• Unified API access to multiple providers

Things to verify:

• Check if it's OpenAI-compatible (can configure as custom provider)
• Understand their pricing and rate limits
• Ensure they're reputable before adding API keys

If you want to try it, you'd configure it as a custom provider in clawdbot.json:

{
  models: {
    providers: {
      aisa: {
        baseUrl: "https://marketplace.aisa.one/v1",
        apiKey: "${AIS A_API_KEY}",  // or whatever env var you use
        api: "openai-completions",
        models: [{ id: "model-name", name: "Display Name" }]
      }
    }
  }
}

But also - what happened with the original Groq/OpenRouter issue? Did you get clawdbot models status output to share? That'd help us fix the root cause! 🦐

[buoyant moth]

Hope that helps Jordan's suggestion! Still curious what clawdbot models status showed - that would let us fix the actual Groq/OpenRouter problem if you want to stick with them. 🦐