UK company intelligence for AI product teams
DataGardener MCP gives software teams, fintechs, consultancies and vertical AI products a remote Model Context Protocol server with read-only, entitled, metered access to connected UK company intelligence. DataGardener operates source ingestion, normalisation, entity resolution and daily updates. Your team builds the experience.
Trust proof for ai and agent builders buyers
Logos indicate organisations using the DataGardener platform, as shown on datagardener.com. They do not indicate use of DataGardener MCP.
The problem
Companies House, iXBRL accounts, charges, PSC, Contracts Finder, Land Registry and spend data all change daily and all have quirks.
Matching an award notice, a payment and a filing to the same company is where most in-house projects stall.
Raw APIs return fields. Agents need tools that express a task, with sensible defaults and provenance.
Example questions
Intelligence layers needed
Example MCP workflow
Add the DataGardener MCP endpoint to your MCP-capable client or framework.
OAuth authorisation binds the connection to a DataGardener entitlement.
Your agent calls task-level tools; the knowledge base tool returns correct field patterns first.
Structured results with source, date and usage are rendered in your own interface.
Choose a question
Illustrative output. Real tool names.
Question
Find companies matching this profile and return structured intelligence for my application.
Stage 1 of 5
The agent breaks the question into the conditions it must satisfy.
Each condition maps to one or more DataGardener intelligence layers. Nothing is guessed: if a condition cannot be evidenced, the agent says so.
Question: Find companies matching this profile and return structured intelligence for my application.
Intent: Profile from my app; Structured output; Entitled modules only; Usage reported.
Tools selected: query_knowledge_base, execute_query, search_ukgaap.
Modules queried: Knowledge base (Fetch field names and query pattern for the profile: Pattern); Company index (Match profile filters, return identifiers and status: Matches); UK GAAP financials (Attach latest financials per match: Enriched).
Evidence connected: Identifiers and status; Latest financial record; Usage count returned with the result.
Answer: Your application receives structured records with registration numbers, status, latest financials and the source and date of each field, plus the usage consumed, ready to render in your own interface.
Provenance: Source Companies House; filed accounts; DataGardener indicators; date Current; module Company, Financial, Risk; usage Reported per call.
Suggested next questions: Add contacts for the matches / Watch these for changes / Page through the next 50
Every answer arrives with its provenance
Time-saving logic
Compare the cost of building and running a UK company-data pipeline with the cost of consuming an operated intelligence layer. The ROI calculator can model analyst time; your engineering estimate covers the rest.
Estimated outcome
This is an estimate computed only from the assumptions you enter, using 4.33 weeks per month and 1,650 working hours per full-time equivalent. It does not include DataGardener MCP costs and is not a guarantee of savings. Validate the acceleration share with a pilot before relying on it.
Relevant proof
UK fintech (unnamed) · Fintech
A UK fintech integrated the DataGardener API for automated onboarding checks across live UK companies with daily refresh.
UK SaaS company (unnamed) · Software
A UK SaaS company used DataGardener targeting to increase demo bookings within ten weeks.
Yes. DataGardener MCP is a remote Model Context Protocol server with OAuth authorisation. Identity travels with the connection.
DataGardener MCP has been used with Claude through Connectors. Any MCP client that supports remote servers with OAuth can connect; other clients are not yet individually tested by DataGardener.
Yes. DataGardener offers a Company Data API for deterministic integrations. MCP is for agents that decide which tools to call; the API is for fixed request and response flows. The MCP vs API guide explains when to use each.
Record-returning tools are metered per company record returned. Describe, count and knowledge-base tools are designed to be used freely before a larger pull. Commercial values are agreed per customer.
Built on DataGardener intelligence already used by leading UK organisations.