Check attached
#Clawdbot update never came back up
fierce pilot
trim pondBOT
Check attached
The image shows Clawdbot failing with ECONNREFUSED 127.0.0.1:11434 — that's your local Ollama server not running. 🦐
Quick fix: Start Ollama first, then Clawdbot:
ollama serve
If Ollama isn't installed:
curl -fsSL https://ollama.ai/install.sh | sh
Or if you meant to use a cloud model instead, check your ~/.clawdbot/clawdbot.json config and make sure your API keys are set.
Let me know if that gets you unstuck! 🦐
fierce pilot
Why do we need Ollama here.
trim pondBOT
Why do we need Ollama here.
You don't need Ollama! The error just means your config is pointing to it. 🦐
You have options:
-
Use cloud models instead (recommended for most — local models need serious hardware):
- OpenAI, Anthropic, Google, etc.
- Just add your API key to
~/.clawdbot/clawdbot.json
-
Run Ollama if you want local models:
ollama serve
To switch to a cloud model, check the docs:
→ https://docs.clawd.bot/concepts/models
What model were you hoping to use? I can help you configure it! 🦐
fierce pilot
All I asked is update it and restart
trim pondBOT
All I asked is update it and restart
Got it, sorry for the detour! 🦐
Can you tell me what command you ran? clawdbot update? clawdbot restart?
Once I know what you ran, I can help you get it back up properly!
fierce pilot
I am already pointed cloud models only
trim pondBOT
I am already pointed cloud models only
Ah, then something in your config is still pointing to Ollama. 🦐
Can you run this and share the output?
clawdbot --version
And then check your config for any Ollama references:
grep -n "ollama" ~/.clawdbot/clawdbot.json
That'll help me spot where the local model is still being called!
fierce pilot
I never used llama
trim pondBOT
I never used llama
Weird — if you never used Ollama, the update might have changed something. 🦐
Can you share (or DM me) the relevant parts of your config? Look for anything under models or agents:
clawdbot gateway call config.get
Redact any API keys, but I want to see what model config you're running. That should tell us where the Ollama connection is coming from!
fierce pilot
It’s been working correctly before i asked it update itself
trim pondBOT
It’s been working correctly before i asked it update itself
That suggests the update pulled in a change that broke something. 🦐
What update command did you run?
clawdbot update?clawdbot gateway call update.run?- Something else?
And what version were you on before vs. now?
clawdbot --version
If it's a recent change that broke things, we might need to:
- Roll back to the previous version
- Or check what changed in the commit
Let me know what command you ran and I'll help you get back up! 🦐
fierce pilot
What’s the best way to ask it update itself
trim pondBOT
What’s the best way to ask it update itself
The standard way:
clawdbot update
Or via the gateway:
clawdbot gateway call update.run
Before you run it again, can you share what version you're on now?
clawdbot --version
And what version worked before? If the update broke something, we can either:
- Wait for a fix
- Roll back to the previous version
Let me know your current version and I'll help you figure out the best path forward! 🦐
fierce pilot
I mean if the bot has to update itself
trim pondBOT
I mean if the bot has to update itself
Ah, got it! 🦐
Clawdbot doesn't have a built-in "update yourself" chat command yet. The standard way is via CLI:
clawdbot update
Or if it's running as a service/systemd:
systemctl restart clawdbot
Why no chat command? It's a safety thing — auto-updating from chat could be a security risk if someone hijacks the conversation.
If you want that feature, you'd need to: