Error: error creating provider: model "llama3.2" not found in model_list

[lynnchin]

I need some help. I keep getting such error message and can't find what's wrong with it.

Here is the config.json:

"agents": {
"defaults": {
"workspace": "~/.picoclaw/workspace",
"restrict_to_workspace": true,
"provider": "ollama",
"model": "llama3.2",
"max_tokens": 8192,
"max_tool_iterations": 50
}
},

Here is what it is shown in the terminal in ollama list:

NAME ID SIZE MODIFIED
tinyllama:latest 2644915ede35 637 MB 38 minutes ago
llama3.2:latest a80c4f17acd5 2.0 GB About an hour ago
tomng/nanbeige4.1:latest dd2a8ce0f368 4.2 GB 10 hours ago

./picoclaw status

Workspace: /home/stanley/.picoclaw/workspace ✓
Model: llama3.2:latest
OpenRouter API: not set
Anthropic API: not set
OpenAI API: ✓
Gemini API: not set
Zhipu API: not set
Qwen API: not set
Groq API: not set
Moonshot API: not set
DeepSeek API: not set
VolcEngine API: not set
Nvidia API: not set
vLLM/Local: not set
Ollama: ✓ http://localhost:11434

Please provide some hints how to get the picoclaw working. Thanks!

[nikolasdehor]

The issue is a mismatch between how you specify the model name in defaults and how PicoClaw resolves it.

Your config has:

• "provider": "ollama" and "model": "llama3.2" in agents.defaults

PicoClaw flow in pkg/providers/legacy_provider.go:19:

1. model = cfg.Agents.Defaults.GetModelName() returns "llama3.2"
2. Since you have no model_list entries, it tries auto-migration from providers config via ConvertProvidersToModelList
3. The Ollama migration in pkg/config/migration.go:210-226 creates a model_list entry with ModelName: "ollama" and Model: "ollama/llama3" (hardcoded)
4. cfg.GetModelConfig("llama3.2") searches for a model_list entry where ModelName == "llama3.2" -- finds nothing
5. Error: model "llama3.2" not found in model_list

The "provider" field in defaults is a legacy hint that only affects the old resolveProviderSelection path. The new model_list path ignores it entirely.

Fix -- use model_list directly:

{
  "model_list": [
    {
      "model_name": "llama3.2",
      "model": "ollama/llama3.2",
      "api_base": "http://localhost:11434/v1"
    }
  ],
  "agents": {
    "defaults": {
      "workspace": "~/.picoclaw/workspace",
      "restrict_to_workspace": true,
      "model": "llama3.2",
      "max_tokens": 8192,
      "max_tool_iterations": 50
    }
  }
}

Key points:

• model_name: "llama3.2" must match the model value in agents.defaults
• model: "ollama/llama3.2" uses the ollama/ protocol prefix so ExtractProtocol() in factory_provider.go routes to the Ollama API base
• api_base is set to the Ollama default. No api_key needed since Ollama local does not require auth
• Remove the "provider": "ollama" field from defaults -- it is ignored in the model_list path and can cause confusion

The picoclaw status output confirms your Ollama is running at http://localhost:11434, so this config should work immediately.

[lynnchin]

Thanks @nikolasdehor! Seems that working now. However, I am getting a TimeoutError in the logs.

2026/03/05 19:21:16 [2026-03-05T11:21:16Z] [INFO] agent: Created implicit main agent (no agents.list configured)
2026/03/05 19:21:16 [2026-03-05T11:21:16Z] [INFO] agent: Agent initialized {tools_count=14, skills_total=6, skills_available=6}
2026/03/05 19:21:16 [2026-03-05T11:21:16Z] [INFO] agent: Processing message from cli:cron: test {session_key=cli:default, channel=cli, chat_id=direct, sender_id=cron}
2026/03/05 19:21:16 [2026-03-05T11:21:16Z] [INFO] agent: Routed message {agent_id=main, session_key=agent:main:main, matched_by=default}
2026/03/05 19:23:16 [2026-03-05T11:23:16Z] [WARN] agent: Timeout error, retrying after backoff {retry=0, backoff=5s, error=failed to send request: Post "http://localhost:11434/v1/chat/completions": context deadline exceeded (Client.Timeout exceeded while awaiting headers)}

I have checked ollama is running in the system and here is the ollama list

NAME ID SIZE MODIFIED
llama3.2:latest a80c4f17acd5 2.0 GB 3 days ago
tomng/nanbeige4.1:latest dd2a8ce0f368 4.2 GB 4 days ago

Tested with ollama run, it works as expected.

ollama run llama3.2 "Test"

It looks like you're testing if I'm working properly. Is there something I can help you with?

Any further insights you can help provide? Cheers.