struggling to get ollama working with openclaw | Friends of the Crustacean 🦞🤝 | Page 1

delicate ledgeBOT Feb 1, 2026, 5:34 AM

#

This is a technical support question - let me help point you in the right direction! 🦐

For Ollama not showing up in the provider list, here are the most common culprits:

Quick checks:

Is Ollama running? ollama list should show your models
Is the port correct? Default is 11434 - make sure it matches your config
Version matters! Run ollama --version - some features need newer versions

Config format reminder:

{
  "openclaude": {
    "providers": {
      "ollama": {
        "host": "http://localhost:11434",

#

        "model": "llama3.2"  // or whatever model you're using
      }
    }
  }
}

If you're still stuck, drop this in #help (channel: 1456457255208878100) with:

Your openclaude config (sanitized)
Output of ollama list
Any error messages you're seeing

The community there has hands-on experience with Ollama setups and can dig deeper into what's not showing up! 🎉

https://docs.openclau.de/providers/ollama for the official docs.

clever obsidian Feb 1, 2026, 12:41 PM

#

ollama with openclaw don't respond why ? how debug please

frosty skiff Feb 1, 2026, 1:25 PM

#

clever obsidian ollama with openclaw don't respond why ? how debug please

me too, stuck with this.
if you know solution, pls share

clever obsidian Feb 1, 2026, 1:26 PM

#

This is my config files :

⁨⁨```
{
"messages": {
"ackReactionScope": "group-mentions"
},
"models": {
"providers": {
"ollama": {
"baseUrl": "http://127.0.0.1:11434",
"apiKey": "ollama-local",
"api": "openai-completions",
"models": [
{
"id": "mistral:7b",
"name": "Mistral 7B",
"api": "openai-completions",
"input": [
"text"
],
"contextWindow": 128000,
"maxTokens": 32000,
"reasoning": false,
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
}
}
]
}
}
},
"agents": {
"defaults": {
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
},
"compaction": {
"mode": "safeguard"
},
"workspace": "/home/stagiaire/.openclaw/workspace",
"model": {
"primary": "ollama/mistral:7b"
}
}
},
"gateway": {
"mode": "local",
"auth": {
"mode": "token",
"token": "9320e03c35ace278f08b15dc59e1d9281841b494b01c7fdf"
},
"port": 18789,
"bind": "loopback",
"tailscale": {
"mode": "off",
"resetOnExit": false
}
},
"skills": {
"install": {
"nodeManager": "npm"
}
},
"hooks": {
"internal": {
"enabled": true,
"entries": {
"boot-md": {
"enabled": true
},
"command-logger": {
"enabled": true
},
"session-memory": {
"enabled": true
}
}
}
},
"wizard": {
"lastRunAt": "2026-02-01T13:22:42.751Z",
"lastRunVersion": "2026.1.30",
"lastRunCommand": "doctor",
"lastRunMode": "local"
},
"meta": {
"lastTouchedVersion": "2026.1.30",
"lastTouchedAt": "2026-02-01T13:22:42.754Z"
}
}

#

and the error message i got on ollama logs :

when i try to talk to him :/

#

i change this :
⁨⁨```
"baseUrl": "http://127.0.0.1:11434",

to 
⁨⁨```
        "baseUrl": "http://127.0.0.1:11434\v1",
```⁩

And create a new model with 32k contexte : 


⁨```
# Créer un Modelfile avec contexte étendu
cat > /tmp/Modelfile << 'EOF'
FROM mistral:7b
PARAMETER num_ctx 32768
EOF

# Créer un nouveau modèle avec contexte 32k
ollama create mistral-32k -f /tmp/Modelfile

# Tester
ollama run mistral-32k "test"

```⁩
and change on ~/.openclaw/openclaw.json : 

⁨```
"models": [
  {
    "id": "mistral-32k",
    "contextWindow": 32768,
    "maxTokens": 16000
  }
]
```⁩
and add ; 
⁨```
"model": {
  "primary": "ollama/mistral-32k"
}
```⁩
restart openclaw gatewway 
⁩

et now i got this error on ollama : 
⁨```
time=2026-02-01T14:32:30.156+01:00 level=INFO source=server.go:1385 msg="llama runner started in 1.53 seconds"
[GIN] 2026/02/01 - 14:32:54 | 500 | 25.821584723s |       127.0.0.1 | POST     "/v1/chat/completions"
time=2026-02-01T14:34:07.339+01:00 level=INFO source=runner.go:682 msg="aborting completion request due to client closing the connection"
[GIN] 2026/02/01 - 14:34:07 | 500 |  59.59742248s |       127.0.0.1 | POST     "/v1/chat/completions"
```⁩

#

but still stuck, i m started to think the problem is my serveur he is not powerfil in memory and cpu to works with openclaw

clever obsidian Feb 1, 2026, 2:12 PM

#

get me crazy no respons from him

mighty horizon Feb 1, 2026, 7:22 PM

#

I have similar problems. Ollama seems to load the machine and just showing "typing" animation, but no respond. After sometime it prompts "Request timed out."

clever obsidian Feb 2, 2026, 11:14 AM

#

Maybe problem come from context

The blog post told 64k minimum context

mighty horizon Feb 2, 2026, 7:00 PM

#

Hmm, im getting somewhere. New version of ollama (0.15.4) seems to be fixed the clawdbot-syntax dilemma, and openclaw start with ollama connection. Right after I promtp a text to the model, and model starts and my PC crashes. I can reproduce the problem...

clever obsidian Feb 2, 2026, 7:06 PM

#

#

Hello i finally make works, but i got a biggg timer problem

#

u can see the same request one with ollama directly and the same with openclaw

#

openclaw he iss tttooooo long why ?

#

same modéle

#

gpt-oss:20b

#

i got respons but it's soooo long why ?

mighty horizon Feb 2, 2026, 7:11 PM

#

Yeap, it seems that im having similar problems, but still i havent had any real text response yet. Model/openclaw instance keeps "typing" OR my PC crashes. Crash might be some other problems i admit...

gloomy plover Feb 2, 2026, 7:11 PM

#

mine doesnt even works man
For one Hi prompt it goes on making 20+ requests for some random bs

#

I am crashing out

#

mistral was fine but the generation time is way too much
and qwen just straight up throwing random bs

mighty horizon Feb 2, 2026, 7:13 PM

#

Which mistral model u were using? Maybe I try that one

gloomy plover Feb 2, 2026, 7:14 PM

#

mistral-nemo:12b

#

what is it even doing 😭 I only sent a Hi

#

bunch of tts responses like wtf

mighty horizon Feb 2, 2026, 7:18 PM

#

It is in havoc...

gloomy plover Feb 2, 2026, 7:19 PM

#

straight up spamming now

#

IS IT EVEN POSSIBLE TO FK IT UP THIS MUCH 😭😭😭😭😭

mighty horizon Feb 2, 2026, 7:20 PM

#

hey atleast you get some reponse. Mine looks like this:

#

clever obsidian Feb 2, 2026, 7:20 PM

#

well openclaw is not optimise for local model, sad

gloomy plover Feb 2, 2026, 7:21 PM

#

😭😭😭😭

clever obsidian Feb 2, 2026, 7:21 PM

#

i think we have to create a bug to get an optimisation of openclaw for local model

#

canu imagine get thislocalyyyyyyyyyyyy would be insaneeeeeeeeee

gloomy plover Feb 2, 2026, 7:21 PM

#

FRR

#

😭THEY SHOULD ATLEAST ADD POLLINATIONS SUPPORT

mighty horizon Feb 2, 2026, 7:21 PM

#

@delicate ledge HELP

delicate ledgeBOT Feb 2, 2026, 7:22 PM

#

mighty horizon <@1457407575476801641> HELP

⚠️ Agent failed before reply: No API key found for provider "anthropic". Auth store: /srv/clawdbot/state/agents/main/agent/auth-profiles.json (agentDir: /srv/clawdbot/state/agents/main/agent). Configure auth for this agent (clawdbot agents add <id>) or copy auth-profiles.json from the main agentDir.
Logs: clawdbot logs --follow

gloomy plover Feb 2, 2026, 7:22 PM

#

lmao

#

cant believe ts

#

I was getting the same error few hours ago

#

and auth-profiles.json didnt even existed

#

Mistral

clever obsidian Feb 2, 2026, 7:24 PM

#

well i have tried openclaw with cloud model and questionme avoir saoul, user.md and many config files, but with this low model he get nothink

#

😢

gloomy plover Feb 2, 2026, 7:24 PM

#

we are so done

#

gotta find a free cloud model

clever obsidian Feb 2, 2026, 7:24 PM

#

openclaw neeed optimisation and would be perfect 😢

gloomy plover Feb 2, 2026, 7:25 PM

#

😭😭😭MISTRAL IS GIVING SOME HOPE

clever obsidian Feb 2, 2026, 7:26 PM

#

i will try ollama cloud model, i think the best one is : glm-4.7:cloud, when u poor no choice 😢

gloomy plover Feb 2, 2026, 7:26 PM

#

rate limits

#

😔

clever obsidian Feb 2, 2026, 7:26 PM

#

yes but even if u can give 200 € to Claude by month 😢

#

no choise for start play with openclaw

#

The creator of openclaw neeed to found money to make a betttttttttttter aps killer, this openclaw is the ia kill apppps, but neeed more optimisation and security

#

all scenerio i see on internet, i so wannnt try them

mighty horizon Feb 2, 2026, 7:28 PM

#

olala i got response, but it took 20 minutes (ollama/mistral:7b)