MCP server

Endpoint: https://api.boogiedata.com/mcp/api (Streamable HTTP, stateless JSON-RPC). Public reads need no key; protected calls take the same X-API-KEY header as REST. These hosts open at launch; until then the configurations below are reference and have not been run against the live host.

Claude Code

claude mcp add --transport http boogie-published https://api.boogiedata.com/mcp/api

Claude Desktop and claude.ai

Add a custom connector with the URL https://api.boogiedata.com/mcp/api.

Cursor

~/.cursor/mcp.json:

{
  "mcpServers": {
    "boogie-published": {
      "url": "https://api.boogiedata.com/mcp/api"
    }
  }
}

VS Code

.vscode/mcp.json:

{
  "servers": {
    "boogie-published": {
      "type": "http",
      "url": "https://api.boogiedata.com/mcp/api"
    }
  }
}

Codex

~/.codex/config.toml:

[mcp_servers.boogie-published]
url = "https://api.boogiedata.com/mcp/api"

OpenAI API

A tool item for the request's tools array:

{
  "type": "mcp",
  "server_label": "boogie_published",
  "server_url": "https://api.boogiedata.com/mcp/api",
  "allowed_tools": [
    "list_datasets",
    "describe_dataset",
    "find_slices",
    "get_coverage",
    "get_sample",
    "list_reports",
    "get_report"
  ],
  "require_approval": "always"
}

Tools

Each tool mirrors one REST endpoint, with the same limits and metering. Read tools are marked read-only.

• list_datasets: GET /v1/datasets
• search_datasets: GET /v1/datasets
• describe_dataset: GET /v1/datasets/{dataset}
• find_slices: GET /v1/datasets/{dataset}/slices
• get_coverage: GET /v1/coverage
• get_sample: GET /v1/datasets/{dataset}/sample
• query_rows: GET /v1/datasets/{dataset}/rows
• get_prices: GET /v1/prices
• list_reports: GET /v1/reports
• get_report: GET /v1/reports/{dataset}/{slice}/latest
• collectible_value: GET /v1/free/collectibles/value (opens with the API)
• venue_fees: GET /v1/free/fees
• fee_venues: GET /v1/free/fees/venues
• str_towns: GET /v1/free/str/towns (opens with the API)
• str_town_check: GET /v1/free/str/town/{slug} (opens with the API)
• account: GET /v1/account
• send_feedback: POST /v1/feedback

Full agent playbook: /agents.md.