Openclaw - your free open-source AI personal assistant (formerly Clawdbot & Moltbot)

[marcusquinn]

@LoudLemur has its own thread:

• https://forum.cloudron.io/topic/14958/agent-zero-a-personal-organic-agentic-framework-that-grows-and-learns-with-you

[robi]

Some tips for setting it up to avoid the dumb clanker feeling:
https://www.notion.so/ashenone/The-Bare-Minimum-To-Make-Your-OpenClaw-Agent-Actually-Good-311664f6a1e381c4976bdb8d689d2ffb

https://xcancel.com/ashen_one/status/2026383178290569539

YT vid didn't load so use ^ - should be fine now.

[andreasdueren]

Packaged, needs more testing though https://git.due.ren/andreas/openclaw-cloudron/-/raw/main/CloudronVersions.json

Plus it's an opinionated packaged geared towards what I think is important but installing new skills should work as is.

[timconsidine]

@andreasdueren well done and thank you

I'm struggling to understand how openclaw being bundled with ollama affects (can be used with) the cloudron ollama app. But finding the answer is deferred below other projects.

[andreasdueren]

@timconsidine I didn't bundle it with ollama. Is that something you want to do? Ollama won't really be very useful on the vast majority of servers without some beefy GPU. But you could connect openclaw with it.

[timconsidine]

@andreasdueren I already have Ollama installed as cloudron app.
I mostly use the ollama cloud models in the cloudron app (ollama pull) so minimal model processing on my cloudron vps.
Then I connect to my cloudron ollama in an AI app.

I'm just not sure what my cloudron ollama is doing with openclaw ! Need to dig into it.

I started packaging openclaw, but paused, might use yours, but generally looking at nanobot for lightweight stuff and AgentZero / Windmill / Dify for heavy weight stuff. So not sure I actually need Openclaw, especially as AgentZero now has skills.
All to be dug into.

[andreasdueren]

Pushing an update now for easy ollama configuration. You should be able to pass endpoint and API key through the .env

To connect your Cloudron Ollama with OpenClaw:

• install Ollama
• download a model in Ollama, for example by running ollama pull tinyllama in the web terminal.
• grab the Ollama API key with cat /app/data/.api_keyor from the file explorer
• install OpenClaw
• open file explorer for OpenClaw app and open .env
• add endpoint and api key from ollama
• restart OpenClaw

[andreasdueren]

actually looks a bit more complicated and that ollama won't let openclaw do model discovery without api key… I'll have to look more into this

[andreasdueren]

I think it works now. Too slow on CPU so can't actually test it without waiting ages. Ollama will only work with a model that supports tool use, so no tinyllama.

1. Node.js version override

Cloudron base image ships Node 22.14, OpenClaw requires 22.16+. We symlink /usr/local/node-22.14.0/bin/node → /usr/bin/node (NodeSource 22.22) so both the gateway
and openclaw CLI use the correct version.

2. Gateway config for reverse proxy

OpenClaw doesn't expect to run behind a proxy by default. start.sh creates /app/data/config/openclaw.json on first run with:

• gateway.mode: "local" — skips pairing requirement
• gateway.trustedProxies: ["172.18.0.0/16"] — trusts Cloudron's Docker network
• gateway.controlUi.dangerouslyAllowHostHeaderOriginFallback: true — allows web UI behind Cloudron's reverse proxy
• gateway.controlUi.dangerouslyDisableDeviceAuth: true — skips device pairing for web UI

3. Ollama auth proxy

OpenClaw's Ollama model discovery (/api/tags) doesn't send auth headers. Cloudron's Ollama package requires Bearer token auth. start.sh starts a Node.js HTTP proxy on localhost:11434 that:

• Intercepts all requests to Ollama
• Injects Authorization: Bearer <OLLAMA_API_KEY> header
• Forwards to the real OLLAMA_BASE_URL
• Overrides OLLAMA_BASE_URL to http://127.0.0.1:11434 so OpenClaw talks to the proxy

Activated when both OLLAMA_API_KEY and OLLAMA_BASE_URL are set in /app/data/.env.

4. Auth auto-provisioning

start.sh auto-creates /app/data/agents/main/agent/auth-profiles.json from:

• CLAUDE_SETUP_TOKEN → type: "token" profile
• ANTHROPIC_API_KEY → type: "api_key" profile

Not sure if this is actually valuable for anyone but it should work.

[robi]

@timconsidine I didn't know ollama could use cloud models.. now it just needs an autorouter feature

@andreasdueren nice is all that for a separate openclaw or the one you can install along with ollama?

[andreasdueren]

@robi These were patches made to the openclaw package. I'd say best is you guys test what you are trying to do then we can make all moving parts work