Weights & Biases MCP

The official Weights & Biases MCP server for W&B Models and Weave, letting AI assistants query and analyze experiment tracking data in natural language. Agents can look up runs and their metrics, compare sweeps, inspect registered model artifacts, and dig into Weave traces to debug LLM application behavior.

{
  "mcpServers": {
    "wandb": {
      "command": "npx",
      "args": [],
      "env": {
        "WANDB_API_KEY": "<YOUR_WANDB_API_KEY>"
      }
    }
  }
}

{
  "type": "object",
  "required": [
    "entity",
    "project"
  ],
  "properties": {
    "entity": {
      "type": "string",
      "description": "W&B entity (user or team)"
    },
    "filters": {
      "type": "object",
      "description": "Optional MongoDB-style filter on run fields"
    },
    "project": {
      "type": "string",
      "description": "Project name"
    }
  }
}

{
  "type": "object",
  "required": [
    "project"
  ],
  "properties": {
    "limit": {
      "type": "integer",
      "description": "Maximum number of traces to return"
    },
    "project": {
      "type": "string",
      "description": "Weave project identifier"
    }
  }
}