tools-only LLM Setup
This page covers the minimum LLM configuration required to run bearlike/Assistant.
Minimum configuration
Set these keys in configs/app.json:
{
"llm": {
"api_base": "https://your-llm-endpoint/v1",
"api_key": "sk-your-key",
"default_model": "gpt-5.2"
}
}
See the optional configuration table below.
Optional LLM configuration
| Key | Purpose | Notes |
|---|---|---|
llm.action_plan_model |
Model for plan generation. | Falls back to llm.default_model if unset. |
llm.tool_model |
Model for tool execution. | Falls back to llm.action_plan_model, then llm.default_model. |
llm.reasoning_effort |
Default reasoning effort level. | Values: low, medium, high, none. |
llm.reasoning_effort_models |
Allowlist for reasoning effort. | Supports exact matches and * suffix wildcards. |
Short walkthrough
- Copy the example config:
cp configs/app.example.json configs/app.json
- Edit the
llmblock with the API base, API key, and model names. - Start a client (CLI, API, or chat). See the client pages for run commands.
MCP setup
MCP servers are optional. When enabled, they add external tools to the registry.
- Create
configs/mcp.json(or run/mcp initin the CLI). - Add MCP server URLs and headers.
- Start a client once to auto-discover tools and cache the manifest under
~/.meeseeks/.
For more details, see Installation.
LiteLLM provider support
The LLM layer is backed by LiteLLM via langchain-litellm.
- Model names can include provider prefixes (for example,
openai/gpt-4-turbo,anthropic/claude-3-sonnet, ormistral/mistral-small). - If
llm.api_baseis set and a model has no provider prefix, the system defaults toopenai/<model>to match OpenAI-style endpoints.