Connect your agent to HubbleField.
Two flows. Both run inside your agent. Search published private-market listings, or publish your company profile - without leaving your agent.
Pick your client, copy the install prompt, and let your agent handle the MCP config, OAuth, and first-query verification.
Let your agent collect company info, create the draft listing, attach documents, and submit for review - without touching the dashboard.
Every agent flow runs over the same open API surface. HubbleField publishes a public MCP server card, OAuth 2.1 discovery metadata, and dynamic client registration so hosted clients can onboard without bespoke connector work. The REST mirror lets you hit the same contracts with plain curl.
What is the HubbleField Connect page for?
Connect is the single hub for agent builders. It covers two flows: searching HubbleField listings from an agent (read scope), and publishing a company profile to HubbleField from an agent (write_listing scope).
Do I need to read the full docs before connecting HubbleField to my agent?
No. Start with the install prompt for your client. The prompts walk your agent through OAuth, config, and verification. Use the full docs only when you need protocol details, REST mirrors, or OAuth registration specifics.
Which agent clients can use HubbleField today?
HubbleField publishes prompts for Claude Desktop, Claude Code, ChatGPT Desktop (readiness), Cursor, OpenClaw, and a generic OAuth MCP client path.
What is the difference between the search and publish flows?
Search uses the read scope and lets an agent query published listings. Publish uses the write_listing scope and lets a founder's agent create a draft listing, attach documents, and submit for review - all without touching the dashboard.
What happens after my agent submits a listing for review?
The listing enters a human review queue. HubbleField staff check it against the publishing rubric and approve or return it within 1 business day. You can track status at your dashboard.
What authentication does HubbleField use?
Hosted clients use OAuth 2.1 with PKCE and dynamic client registration. Local stdio clients can use manual bearer tokens minted after email verification. The search flow uses the read scope; the publish flow requires write_listing.