Gmail integration for AI agents.
Let your AI agents work in Gmail — read threads, draft and send email, and search the inbox — through a brokered, audited connection. Open Connector runs the OAuth, seals the token in an encrypted vault, and serves Gmail tools to your agent over MCP or a typed API — credentials injected server-side, every call audited, nothing leaving your infrastructure. Open source (AGPL-3.0) and self-hostable.
Real Gmail actions, brokered and audited.
Your user connects Gmail once; your agent can then read threads, draft and send email, and search the inbox — scoped to the OAuth permissions you grant and the tool allowlist you configure. Every action is least-privilege and written to a tamper-evident audit trail.
- 1
Your user grants Gmail access once (OAuth) — the token lands in the vault.
- 2
Your agent calls a tool over MCP or the typed API; Open Connector injects the credential server-side.
- 3
Every brokered call appends a hash-chained audit record — nothing leaves your infra.
Gmail integration, answered
- How do AI agents use Gmail through Open Connector?
- Your user connects Gmail once (OAuth). Open Connector stores the token in an encrypted vault and exposes Gmail tools to your agent over MCP or a typed API, injecting the credential server-side on each call — the agent never sees a raw token, and every action is audited.
- Is this a Gmail MCP server?
- Yes. Open Connector can serve Gmail as a named MCP server with a scoped tool allowlist and a per-user mcp_url, so any MCP client connects and calls Gmail tools with credentials brokered server-side.
- Where do Gmail OAuth tokens live?
- In your own infrastructure. Self-host Open Connector and the Gmail token is sealed in an AES-256-GCM vault in your Postgres, injected at call time — it never leaves your environment.
Give your agents Gmail — keep the keys.
Open source, self-hostable, with Gmail credentials that never leave your infrastructure. Run it from source today.