hi

[woeful lantern]

hi

[last ferry]

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> ✅ Mark Solution

[woeful lantern]

so i have a api key for a local ollama model how do i connect that with openclaw

[shy bay]

hey, what can I help you with? 🦐

Let me verify the Ollama docs so I give you the right config keys, not vibes.

first, paste openclaw --version so I can match the right setup.

If you mean a local Ollama server, the easiest path is:

1. make sure Ollama is running and you have a model pulled
   ollama pull gemma4
2. set the Ollama env var
   • local-only: export OLLAMA_API_KEY="ollama-local"
   • Ollama cloud: export OLLAMA_API_KEY="your-real-key"
3. run: openclaw onboard
4. choose Ollama
5. pick Local only or Cloud + Local
6. verify/set the model
   openclaw models list --provider ollama
openclaw models set ollama/<your-model>

Important: if you use a custom Ollama host, use http://host:11434 with no /v1. The /v1 URL is the OpenAI-compatible path and tool calling gets unreliable there.

If you want manual config instead of onboarding, I can give you the exact JSON5 too.

Docs:

• https://docs.openclaw.ai/providers/ollama
• https://docs.openclaw.ai/help/faq#can-i-use-self-hosted-models-llamacpp-vllm-ollama

If you send:

• openclaw --version
• local vs remote Ollama
• local-only vs cloud+local

I’ll give you the exact next command.

[woeful lantern]

ok but how do i add the thingy and conenct them

[shy bay]

yeah, the “thingy” is just the OLLAMA_API_KEY env var 😄

For local Ollama, do this:

ollama serve
ollama pull gemma4
export OLLAMA_API_KEY="ollama-local"
openclaw onboard

Then in the setup wizard choose:

• Ollama
• Local only
• pick your model

After that, run:

openclaw models list --provider ollama
openclaw models set ollama/gemma4

That’s basically the connection:

• Ollama runs on http://127.0.0.1:11434
• OpenClaw sees OLLAMA_API_KEY
• OpenClaw discovers the Ollama models and uses them

Important:

• for local Ollama, the key does not need to be real, ollama-local is fine
• if OpenClaw is already running in the background, restart it after adding the env var
• if you use a custom Ollama URL, use http://host:11434 and not /v1

Docs: https://docs.openclaw.ai/providers/ollama

If you want, I can give you the exact commands for your OS next.

[woeful lantern]

how do i open openclaw again

[shy bay]

if you mean the OpenClaw web UI, run:

openclaw gateway start
openclaw dashboard

If the gateway is already running, just:

openclaw dashboard

You can also open it directly at: