Skip to main content
Orbb reads the public professional record — shared employers, shared schools, investor roles, LinkedIn comments — and turns it into a graph of who knows whom, with the evidence behind every relationship. You can reach it two ways.

The API

Four calls over HTTPS: resolve a person, list their circle, test an edge between two people, find paths to someone. From a fifth of a cent a call.

The MCP servers

Connect Claude, ChatGPT, Cursor or your own agent. It asks Orbb who knows somebody, plans research, and reads the results, in your agent’s own conversation.

What you can ask

A token is a dollar. Reading a job’s answer and your balance is free. See Billing.

The one rule: an unknown is not a no

Every answer says what was checked, not only what was found. A pair we never examined comes back UNKNOWN, never NO_EVIDENCE_FOUND; a strength that is absent is not a zero. Read Unknown is not no before you build anything that filters people out — it is the mistake the API is designed to stop you making.

Start here

Quickstart

A key, a resolve, a circle. Five minutes.

Building with an agent

llms.txt, Markdown pages, and a docs MCP server for your coding agent.

API reference

Every field, every error code, with a playground.
The API and Orbb Research read the public record only: nothing inside your CRM, mailbox or calls. Strength measures the evidence for a relationship — how much two people overlapped, how directly they interacted — not whether it is warm today. Treat a strong relationship as worth checking with whoever owns it.