Forest Valley · Climate Tech Data Hub
Connect the Data Hub to your AI assistant
How to connect the Data Hub MCP server to Claude, ChatGPT, Cursor and other AI clients. The page is public: you can share it or paste it into your assistant. A Markdown version is at /mcp/setup.md.
Connection details
Forest Valley Data HubRemote MCP, streamable HTTPhttps://data-hub.forestvalley.org/mcpNone: open and read-only120 requests per minute per IP addressAdd a remote HTTP MCP server at https://data-hub.forestvalley.org/mcp. No token or login is needed. Start with list_filters to see the accepted values, then search_startups; get_startup returns the source URLs.
What this gives you
MCP (Model Context Protocol) lets an AI assistant query a data source directly. With this server, the assistant can search 1,247 climate tech startups and 68 funds, accelerators and programmes, open the full record of a startup, and compute counts and funding totals, with the date of the data and the source URLs in each answer. Nothing is installed on your computer, and the server cannot change any data.
Steps by client
- Claude (claude.ai and the desktop app): Settings, Connectors, Add custom connector, paste the endpoint.
- Claude Code: run
claude mcp add --transport http forest-valley-data-hub https://data-hub.forestvalley.org/mcp
- ChatGPT: Settings, Connectors; turn on developer mode if the option is not visible; add a custom MCP server and paste the endpoint.
- Cursor, VS Code and other clients with a JSON configuration: add
{ "mcpServers": { "forest-valley-data-hub": { "url": "https://data-hub.forestvalley.org/mcp" } } }
Menu names change between versions: look for Connectors, Integrations or Custom MCP server. If a client only accepts a local server (stdio), it cannot use this address.
Tools
| Tool | Parameters | What it does |
|---|---|---|
search_startups | query, country, sector, technology, trl_min, last_round_year, last_round_type, limit, offset | Search the Climate Tech Data Hub startups. All filters are optional and combined with AND. Returns summaries with the date of the last update; use get_startup for the full record with source URLs. |
get_startup | id, name | Full record of one startup by id or exact name, with the source URLs of each field, which fields are inferred, and the last update date. |
search_investors | query, type, geography, stage, focus, limit, offset | Search funds, accelerators and programmes behind the startups. |
get_investor | id, name | One fund, accelerator or programme by id or exact name, with the startups it backed. |
list_filters | none | The values accepted by the filters of search_startups and search_investors, with counts. |
market_stats | country, sector, technology | Aggregates over the startups: counts by country, sector, technology, TRL and last-round type, and funding totals. |
Sources and dates
Every answer carries the date the data was generated (2026-11-20) and the date of the latest startup enrichment (2026-11-04). get_startup adds, for each field, the source URLs and the date of the last update of that startup.
Source URLs exist for 238 of the 1,247 startups. Fields marked inferred are syntheses of other fields and have no URL. The update date applies to the whole startup record, not to each field.
Example questions
- Which startups work on direct air capture, and when was each last updated?
- List the Italian startups at TRL 6 or higher in industrial decarbonisation, with their last funding round.
- Give me the full record of one startup with the source URL of each field.
- Which funds and accelerators back startups in energy storage?
- How many startups are in each country, and what is the total funding?
Read more about the Data Hub at ai-lab.forestvalley.org/data-hub; the database is at data-hub.forestvalley.org. Forest Valley Institute Srl Impresa Sociale · Via Donatello 5B, 20131 Milano, Italy · VAT IT11849620965