meta-data-mcp

Name: meta-data-mcp
Author: derek-linz

As of version 2.5, meta-data-mcp now actually works with Claude Desktop.

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

Opendata Providers Find: Search the meta-data-mcp plugin registry. Returns plugins that match a free-text query and/or domain/region filters. Use this FIRST when you don't know which plugin can answer
Opendata Plugins Create: Autonomously create a new plugin for this meta-data-mcp server from a YAML spec. Use this when `opendata_providers_find` returned no match. Recommended flow: first call `opend
Opendata Plugins Draft: Build a validated YAML plugin spec from structured inputs. Use this BEFORE `opendata_plugins_create` to avoid hand-writing YAML. Validates id format, kebab-case tool names, pat

What data it sees

Do you need an account

No: the server works without sign-in

As of version 2.5, meta-data-mcp now actually works with Claude Desktop.

Previously, a harness limitation in Claude prevented hot-add of mcp-servers (the provider plugins).

One MCP server gatekeeping hundreds of tool calls to open data sources across all domains--no API keys required. Ask about GDP, earthquakes, court records, air quality, drug approvals, satellite imagery, or biodiversity — meta-data-mcp routes the request to the right API and returns structured data. Providers activate lazily so the default tool surface stays small; use opendata-find-providers to discover and opendata-activate-provider to unlock domain-specific tools on demand. Run it locally or with streaming sse.

How to connect

How to connect meta-data-mcp to Claude

The server address to paste into Claude settings:

https://meta-data-mcp--derek-linz.run.tools

Open Claude (claude.ai or the desktop app).
Go to Settings → Connectors.
Click Add custom connector.
Paste the server address copied below into Remote MCP server URL and click Add.
In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect meta-data-mcp to ChatGPT

The server address to paste into ChatGPT settings:

https://meta-data-mcp--derek-linz.run.tools

Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
Open Settings → Connectors and click Create.
Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect meta-data-mcp to Cursor

The server address to paste into Cursor settings:

https://meta-data-mcp--derek-linz.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (15)

Raw names from tools/list. Only developers need these.

opendata_providers_find	Search the meta-data-mcp plugin registry. Returns plugins that match a free-text query and/or domain/region filters. Use this FIRST when you don't know which plugin can answer a question. If no plugin matches, the response includes a `next_step` field that explains how to autonomously create one via `opendata_plugins_create`.
opendata_plugins_create	Autonomously create a new plugin for this meta-data-mcp server from a YAML spec. Use this when `opendata_providers_find` returned no match. Recommended flow: first call `opendata_plugins_draft` with structured fields to get a valid YAML spec, then pass it here. The new plugin is materialized to disk, imported, registered in the live registry, and its tools become available immediately. On success this also opens a public contribution PR of the generated plugin to the project so others can use it; set META_DATA_MCP_AUTO_CONTRIBUTE=0 to disable.
opendata_plugins_draft	Build a validated YAML plugin spec from structured inputs. Use this BEFORE `opendata_plugins_create` to avoid hand-writing YAML. Validates id format, kebab-case tool names, path-placeholder/param consistency, parameter types, and response format. Returns the YAML string ready to feed into `opendata_plugins_create`.
opendata_explain_choice	Explain the scoring breakdown for a provider search. Shows how each provider was ranked using token matching, fuzzy matching, semantic similarity, and metadata filters.
opendata_domains_list	List the controlled domain vocabulary used by the provider registry (e.g. 'health', 'legal', 'finance', 'earth-science').
opendata_regions_list	List the controlled region vocabulary used by the provider registry (e.g. 'us', 'eu', 'uk', 'global').
opendata_providers_describe	Fetch the full registry entry for a single provider id — title, description, domains, regions, keywords, homepage, license note, required environment variables.
opendata_providers_list	Enumerate all providers in the opendata-mcp registry (paginated, terse). Returns id, title, domains, regions, and any required env vars per provider.
opendata_providers_activate	Activate a registered provider so its tools become callable in this session. By default the server starts in discovery-only mode — only meta tools (find-providers, list-providers, etc.) are advertised. Activation imports the plugin module and merges its tools into the advertised list, then sends a tools/list_changed notification so the client refetches its catalog. Idempotent.
opendata_providers_deactivate	Remove a previously-activated provider's tools from the session's advertised list. The Python module remains imported (Python caches modules) but its tools no longer appear in tools/list. A tools/list_changed notification is sent so the client refetches.
opendata_providers_list_active	List the providers currently activated in this session, along with the tool names each contributes. Useful for inspecting why a particular tool is (or isn't) advertised.
opendata_health_snapshot	Snapshot the in-memory provider health registry. Returns a score in [0.0, 1.0] for each requested provider (or every registered provider when called without arguments). Health degrades on recent 5xx / 429 / network failures and decays back toward 1.0 over ~5 minutes; 401/403 are excluded as caller misconfig. The discovery app uses this to paint live health badges next to each search result.
opendata_tool_call	Proxy any activated plugin tool by name. Use this when dynamically activated tools aren't directly callable in your environment. First activate the provider with opendata_providers_activate, then call this tool with the tool name and arguments from the activation response's tool_schemas field.
opendata_federate_query	Run several plugin tool calls, normalize their results onto a common geography + time axis, and merge them into one cited series. Auto-activates each query's provider. Use this to overlay the same indicator from different open-data sources (e.g. Eurostat vs World Bank) in a single answer.
opendata_federate_compare	Federate several plugin tool calls and report a coverage matrix — which sources cover which geographies and periods — alongside the merged series. Use this to spot gaps or disagreements between open-data sources.