Find UK seed-stage climate software companies and show source links and freshness notes.
MCP Server for private-company discovery
HubbleField's MCP Server lets an AI agent search and inspect structured public company profiles through account-bound tools. It is built for private-company discovery workflows, not as a decision tool, transaction venue, pricing opinion surface, or route into private diligence.
Information only. Not investment advice. Not an offer of securities in any jurisdiction.
Inspect this placeholder profile and summarise public categories, geography, stage, and update state.
List companies currently marked with public raise context and include their HubbleField profile URLs.
Prepare a research list of workflow automation companies for human review.
Public profile fields
- company name, slug, website, geography, categories, stage, and profile URL
- short public description, founder-approved positioning, and visible source links
- public raise context, access paths, freshness state, and profile update dates
- trust-signal summaries and routes back to stable HubbleField pages
Gated private diligence
- cap table, full financial model, customer lists, contracts, and diligence files
- unpublished listings, internal review notes, identity documents, and personal data
- founder-only dashboards, sensitive raise material, and private forecasts
- account-bound actions unless the signed-in user or token has the required permission
search_companies
Search published public profiles by sector, geography, stage, text query, and public raise context.
get_company
Fetch one public profile by slug so an agent can inspect the same public context a human can read.
get_raising_now
Return public profiles with visible raise context for research workflows that need current raise status.
list_categories
Expose supported categories so agents can guide users toward available filters before searching.
Use the MCP tools reference for exact schemas, resources, prompts, and REST mirrors. Use this page when you need the setup, workflow, and public/private boundary in plain English.
One disclosure boundary, two ways to connect.
REST is the direct HTTP surface for product integrations, scripts, and internal tools. MCP is the agent-native tool surface for Claude, Cursor, ChatGPT-compatible clients, and generic MCP clients that ask tools to act on a user's behalf.
Both surfaces use the same public profile layer until an authenticated, permissioned route grants more. Read the REST/API docs for endpoint detail and the data API page for commercial API review.
Set up in your agent client
- Choose a search setup path on /connect for Claude Desktop, Claude Code, Cursor, OpenClaw, ChatGPT Desktop, or a generic MCP client.
- Use OAuth for hosted clients when available, or a session-minted bearer token for local stdio clients.
- Start with read-only search tools, then move to account-bound workflows only after the user signs in and grants the required scope.
- Keep private documents and founder-only material outside agent responses unless a separate permissioned path explicitly allows access.
Example agent prompt and public-safe response shape.
Use HubbleField to find three UK seed-stage climate software companies. For each result, return the public profile URL, public raise context, categories, freshness state, and source links. Do not include private documents or founder-only data.
{
"query": "UK seed-stage climate software",
"results": [
{
"name": "Example Robotics Ltd",
"slug": "example-robotics",
"profile_url": "https://hubblefield.com/co/example-robotics",
"geography": "United Kingdom",
"stage": "seed",
"categories": ["climate_software", "industrial_workflows"],
"public_raise_context": "Public placeholder raise context, if visible on the profile.",
"freshness_state": "fresh",
"source_links": ["https://example.com/example-robotics"],
"notes": "Synthetic placeholder record for response-shape review."
}
],
"private_fields_omitted": true
}These examples use synthetic placeholder data only. They do not include live company records, bearer tokens, private documents, identity data, or founder-only diligence.
Investor and adviser research
Ask an agent to search structured public profiles, gather profile links, and hand the result to a human for review.
Corporate-development scans
Let an internal agent collect public company context by category, geography, and stage before deeper manual diligence.
Founder visibility
Give founders a public profile layer that AI agents can parse without exposing private documents or unpublished details.
Agent-builder prototypes
Connect MCP or REST to workflow tools that need predictable private-company profile shapes and clear permission limits.
What is HubbleField's MCP Server for?
It gives AI agents a structured way to search public HubbleField company profiles, inspect public context, and route users to setup or account-bound workflows.
How is MCP different from REST/API access?
REST is the direct HTTP surface for product teams and scripts. MCP is the agent-native tool surface for clients that call tools on a user's behalf. Both use the same public and permissioned data boundary.
Does the public MCP surface expose private diligence?
No. Cap tables, full financial models, customer lists, contracts, diligence files, unpublished listings, personal data, and sensitive raise material stay gated.
Which agent clients can use HubbleField?
The connect page provides search setup paths for Claude Desktop, Claude Code, Cursor, OpenClaw, and generic MCP clients. The ChatGPT Desktop prompt checks current MCP support before making changes.
Is this investment advice?
No. HubbleField is information only, not investment advice, and not an offer of securities in any jurisdiction.
Connect the agent surface your team already uses.
Start with the MCP setup guide if you want your AI client to call HubbleField tools. Start with REST if you are building a direct product integration. Join the agent-builder access path if you want launch updates and integration help.