Configuration

Metacoder uses YAML configuration files to configure AI assistants and their extensions.

Configuration File Format

Basic Configuration

ai_model:
  name: model-name
  provider:
    name: provider-name
    api_key: your-api-key

Extensions (MCPs)

You can extend coders with Model Context Protocol (MCP) servers:

ai_model:
  name: gpt-4
  provider: openai

extensions:
  - name: filesystem
    command: npx
    args: [-y, "@modelcontextprotocol/server-filesystem"]
    enabled: true
    type: stdio

  - name: github
    command: uvx
    args: [mcp-github]
    env:
      GITHUB_TOKEN: ${GITHUB_TOKEN}
    enabled: true
    type: stdio

For detailed information about MCP configuration and usage, see the MCP Support documentation.

Using Config Files

metacoder "Your prompt" --config myconfig.yaml

Custom Instructions

You can provide custom instructions to any coder using the --instructions option:

metacoder "Your prompt" --instructions custom_instructions.md

The instructions file is loaded and passed to the coder's instruction handling mechanism. Each coder will use these instructions according to its own configuration:

• Claude: Instructions are written to CLAUDE.md
• Goose: Instructions are written to .goosehints
• Gemini: Instructions are written to GEMINI.md
• Other coders: Check the specific coder documentation

Example instructions file:

# Custom Instructions

You are an expert Python developer following these guidelines:
1. Use type hints for all functions
2. Write comprehensive docstrings
3. Follow PEP 8 style guidelines
4. Include unit tests for all new functions

You can combine instructions with other configuration options:

metacoder "Refactor this code" \
  --coder claude \
  --instructions style_guide.md \
  --config claude_config.yaml \
  --mcp-collection tools.yaml

Environment Variables

Config files support environment variable expansion:

ai_model:
  provider:
    api_key: ${OPENAI_API_KEY}

Per-Coder Defaults

Each coder has default configurations that are automatically applied. Custom configs override these defaults.