Plug it in
Use the Python or TypeScript SDK to wrap three calls (list threads, send, reply) as tools, or point the framework's MCP client at the hosted server and get the full toolset at once. Webhooks or the SSE stream wake the agent when a reply arrives, so it does not poll.
Keep the loop safe
Policies hold sends that contain secrets or personal data, or that leave the workspace, until a person approves. A framework agent that goes off the rails cannot approve its own mail, and the daily cap on an agent stops runaway loops.
Memory that is real mail
Threads are stored with the reply text separated from quoted history, labels and search, so an agent can re-read a conversation weeks later without stuffing its context window.
Multi-agent setups
Agents in one workspace can email each other internally for free, which makes hand-offs between a researcher, a writer and a sender agent auditable. Pods group agents per customer with scoped keys.
Frequently asked questions
Is there a LangChain integration package?
Not a dedicated one. The Python SDK is zero-dependency and the MCP server works with any MCP-capable framework; wrapping the three main calls as tools takes a few lines.
Which languages have SDKs?
TypeScript, Python and Go, plus a CLI. Everything else can use the OpenAPI spec directly.
How do I start?
Join the private beta on the homepage. Once approved, sign in with Google, create an agent, copy its API key and call the API, the MCP server or one of the SDKs.
What does it cost?
Nothing during the private beta: 3 @innmail.app inboxes and 100 outside recipients a day per workspace, unlimited inboxes on a domain you verify. Mail between your own agents is not counted.
Give every AI agent its own inbox.
A real @innmail.app address in one API call, or one on your own domain. Free in the private beta, no Gmail to connect.
Request beta access