What Do They Know
This MCP server lets AI assistants research UK Freedom of Information requests and authorities on WhatDoTheyKnow, and optionally draft or update requests.
Search FOI requests and response events with structured expressions such as status or body filters.
Search UK public authorities, retrieve authority profiles/statistics, and access the full authorities CSV resource.
Fetch request timelines, full request details, correspondence text, and attachment metadata where exposed.
List a user’s visible FOI requests and fetch user activity feeds.
Build prefilled WhatDoTheyKnow request URLs for an authority and topic.
Create request records and update user-assessed request states via the experimental write API when
WDTK_API_KEYis configured.Access MCP resources for authority, request, user, and feed data.
Use the FOI request drafting prompt and generic prompt listing/rendering tools.
whatdotheyknow-mcp
Maintained by BOUCH — private hosting, custom tools and setup: paul@bouch.dev · bouch.dev/products/whatdotheyknow-mcp
A Model Context Protocol server for UK Freedom of Information research. Connects AI assistants to WhatDoTheyKnow — the UK's largest FOI request platform — to search requests, read responses, look up public authorities, and draft new requests.
Tools
Tool | Description |
| Full-text search of FOI requests and responses via WhatDoTheyKnow's Atom feed. Supports structured expressions ( |
| Search UK public authorities by name. Returns slug for use with other tools. |
| Fetch the event timeline (sent, response, clarification) for a specific FOI request. |
| Fetch full public request JSON, including status, authority, requester, correspondence events, and attachment metadata where exposed. |
| List a user's visible requests with title, slug, status, authority, dates, and snippets. |
| Fetch a user's activity Atom feed as structured entries with readable content text. |
| Fetch full public authority JSON, including contact/profile fields and request statistics. |
| Build a prefilled WhatDoTheyKnow request URL for a given authority and topic. |
| Create a request via the write API (requires |
| Update user-assessed state of a request (requires |
Related MCP server: mcp-legislation-uk
Resources
URI template | Returns |
| Authority profile JSON |
| FOI request detail JSON |
| User profile JSON |
| Request event Atom feed |
| User activity Atom feed |
| Full CSV of all UK public authorities |
Prompts
Prompt | Description |
| Draft a narrow, specific FOI request for a given authority and topic. |
Connect
Hosted (no install)
{
"mcpServers": {
"whatdotheyknow": {
"type": "http",
"url": "https://whatdotheyknow-mcp.fly.dev/mcp"
}
}
}Local (uvx)
{
"mcpServers": {
"whatdotheyknow": {
"type": "stdio",
"command": "uvx",
"args": ["whatdotheyknow-mcp"]
}
}
}Environment variables
Variable | Required | Description |
| Optional | Enables |
Upstream API and Licence
Source | API | Licence | Auth |
WhatDoTheyKnow |
| None (read) / API key (write) |
Data is sourced directly from the WhatDoTheyKnow public API. The platform is operated by mySociety.
Available Tools
12 toolsbuild_request_urlARead-onlyIdempotent
Build a prefilled WhatDoTheyKnow request URL.
Inputs: authority_slug (required), body, tags, title, +1 more. Returns: url, authority_slug
create_request_recordADestructive
Create a request through the experimental write API.
Requires WDTK_API_KEY in the server environment.
Inputs: body (required), title (required), external_url (required), external_user_name (required). Returns: object
get_authority_detailARead-onlyIdempotent
Return full public JSON for a WhatDoTheyKnow public authority.
The response includes contact/profile fields, tags, publication links, and request statistics such as successful, overdue, and classified request counts where WhatDoTheyKnow exposes them.
Authority slugs use underscores, not hyphens — e.g. 'home_office' not 'home-office'. Use search_authorities to find the correct slug.
Inputs: authority_slug (required). Returns: object
get_promptA
Get a prompt by name with optional arguments.
Returns the rendered prompt as JSON with a messages array. Arguments should be provided as a dict mapping argument names to values.
Inputs: name (required), arguments. Returns: result
get_request_detailARead-onlyIdempotent
Return full public JSON for a WhatDoTheyKnow FOI request.
The response includes request metadata, public body details, requester details, state/status fields, and the visible info_request_events array containing correspondence text and attachment metadata where WhatDoTheyKnow exposes it.
Inputs: request_slug (required). Returns: object
get_request_feed_itemsARead-onlyIdempotent
Return parsed Atom feed entries for a specific FOI request as structured objects.
Use this instead of reading the raw wdtk://requests/{slug}/feed resource when you
want structured AtomEntry objects rather than raw XML. Each entry's link field
contains the request URL; use the slug from that URL with get_request_detail
for full detail.
Inputs: request_slug (required), limit. Returns: result
get_user_feed_itemsARead-onlyIdempotent
Return parsed Atom feed entries for a user's WhatDoTheyKnow activity.
Unlike the raw wdtk://users/{slug}/feed resource, this returns structured
entries with the Atom content converted to readable text in content and
mirrored to summary for clients that only display summary fields.
Inputs: user_slug (required), limit. Returns: result
get_user_requestsARead-onlyIdempotent
List a user's visible WhatDoTheyKnow requests from their public requests page.
Returns request title, slug, URL, current displayed status, authority, updated
timestamp, event text, and the page snippet. status_filter is a
case-insensitive substring match against the displayed status.
Inputs: user_slug (required), limit, status_filter. Returns: result
list_promptsA
List all available prompts.
Returns JSON with prompt metadata including name, description, and optional arguments.
Inputs: no named parameters. Returns: result
search_authoritiesARead-onlyIdempotent
Search WhatDoTheyKnow public authorities by name.
Returns up to limit authorities whose name or short_name contains query
(case-insensitive). Use the slug field with authority_json or
build_request_url as the next step.
Example: search_authorities("Liverpool") → slug "liverpool_city_council" Then: authority_json with that slug, or build_request_url with it.
Inputs: query (required), limit. Returns: result
search_request_eventsARead-onlyIdempotent
Search WhatDoTheyKnow's feed-based event index and return structured results.
Call this to find FOI requests matching a query expression. Returns up to limit
AtomEntry objects. Use the link field of each result as the next navigation
step — extract the request slug and call get_request_detail or
get_request_feed_items for full detail.
Example expressions: status:successful body:"Liverpool City Council" (variety:sent OR variety:response) status:successful
Inputs: search_expression (required), limit. Returns: result
update_request_stateADestructive
Update the user-assessed state of a request through the experimental write API.
Requires WDTK_API_KEY in the server environment.
Inputs: state (required), request_id (required). Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 12 tools
Most tools target clearly distinct resources and actions: request detail vs. authority detail vs. user requests vs. authority search. The main overlap is between feed/event tools that both return AtomEntry-style results, but descriptions clarify which to use. The prompt tools are unrelated to the FOI domain, though not easily confused with the domain tools.
All tool names use consistent snake_case and follow a predictable verb_noun pattern: get_*, search_*, build_*, create_*, update_*, list_*. No naming convention mixing is present.
Twelve tools is a well-scoped size for this domain, covering reads, search, URL building, and limited writes. The two prompt tools are somewhat extraneous, but the overall count remains reasonable and manageable.
The surface covers the core FOI workflows: request detail/feed, user activity, authority search/detail, event search, URL construction, and experimental create/update operations. Minor gaps remain, such as attachment content retrieval, user profile detail, and richer request lifecycle operations, but agents can work around most of them.
Maintenance
Related MCP Connectors
Hosted MCP server for live public-data APIs and Skills for AI agents.
CompanyLens is a remote MCP server giving AI agents instant access to official company registry data across 19 jurisdictions in Europe, the Americas, and Asia-Pacific. Eighteen read-only tools let you search companies and people, look up officers and beneficial owners, map corporate networks through shared directors, screen names against the UK disqualified directors register, find every company at a registered address, and pull filing history — all from a single connector. Visit our website: https://companylens.io
**UK Source Intelligence — GOV.UK Change Monitoring & Historical Evidence for AI Agents** UK Source Intelligence (UKSI) is an MCP-compatible monitoring service that helps AI agents and developers detect, investigate and verify changes to official GOV.UK content. Unlike a simple website-change notification tool, UKSI retains observed wording, timestamps and SHA-256 fingerprints, allowing agents to compare genuine historical versions and provide traceable evidence to their human users. **Key capabilities:** - Nine pre-monitored GOV.UK pages. - Monitor additional supported GOV.UK pages. - Compare previously observed and updated wording. - Retrieve historical snapshots, timestamps and source fingerprints. - Verify exact quotations against official government content. - Access information through 12 MCP tools or an API. - Free public evidence demonstration without registration. **Free trial — 14 days** - No payment required. - 10 API requests per day. - One custom monitored GOV.UK page. **Introductory access — £4.99 one-time payment** - 30 days of access. - 100 authenticated API requests per day. - Three custom monitored GOV.UK pages. - No subscription or automatic renewal. **Ideal for:** AI developers, compliance researchers, accountants, regulatory monitoring workflows and organisations requiring verifiable historical GOV.UK source evidence. **Limitations:** Monitoring begins when a page is registered; historical records cannot be reconstructed from before monitoring began. Coverage is restricted to supported GOV.UK content, and monitoring intervals are not guaranteed. UKSI is not a substitute for professional legal or regulatory advice. Website: https://uk-source-intelligence.pages.dev/ MCP endpoint: https://uk-source-intelligence-mcp.mazdamen.workers.dev/mcp
Read-only search over 2,065+ AI agent tools and 276 MCP servers. Free, no sign-up needed.
Related MCP Servers
- AlicenseBqualityBmaintenanceAn MCP server that gives AI assistants access to UK Parliament data. Query MPs, Lords, bills, votes, committees, debates, and more through AI assistants like Claude Desktop and VS Code Copilot.2339MIT
- AlicenseNot gradedqualityCmaintenanceAn MCP server for accessing the UK's official legislation database (legislation.gov.uk), enabling natural language queries via Pipeworx's AI gateway.80 npmMIT
- AlicenseNot gradedqualityDmaintenanceMCP server for the Froide FOI platform. Exposes Froide's REST API as MCP tools for Claude and other MCP clients.MIT
- FlicenseNot gradedqualityDmaintenanceMCP server exposing the full UK Companies House Public Data API, enabling natural language queries for company profiles, search, officers, filing history, charges, insolvency, and persons with significant control, as well as downloading and reading PDF documents.-