AgentMailer developers

Build against contracts agents can discover.

Choose hosted MCP for tool-using agents or the versioned REST API for application code. Both preserve the same human-approved identity and permission boundary.

Two supported paths

Use MCP to act or OpenAPI to integrate.

Tool-using agents

Hosted Streamable HTTP MCP

Connect an OAuth-capable client, complete human approval, call auth_me, and request only the permissions needed by the task.

https://api.agentmailer.ai/mcp
Application code

REST and generated clients

Generate a typed client from the stable OpenAPI 3.1 contract or use an official SDK and a human-approved credential.

https://api.agentmailer.ai/openapi.json

Machine discovery

Fetch the contract instead of guessing.

OpenAPI 3.1
https://api.agentmailer.ai/openapi.json
OAuth protected resource
https://api.agentmailer.ai/.well-known/oauth-protected-resource
MCP server card
https://api.agentmailer.ai/.well-known/mcp/server-card.json
MCP registry manifest
https://api.agentmailer.ai/.well-known/mcp.json
Documentation MCP
https://api.agentmailer.ai/docs/mcp
Documentation MCP manifest
https://api.agentmailer.ai/.well-known/mcp/docs.json
Machine guide
https://api.agentmailer.ai/llms.txt

SDKs, CLI and examples

Start from a maintained client surface.

The public AgentMailer repository contains TypeScript, Python, Rust, Ruby, Go, and Swift clients, the CLI, portable skills, and paired Python and TypeScript examples.

Browse AgentMailer developer clients

Safe retries

Build for delivery, replay, and backpressure.

  • Use a stable Idempotency-Key for retried write operations.
  • Honor Retry-After on 429 responses.
  • Read RateLimit-Policy as a shared capacity hint, not a service guarantee or per-caller balance.
  • Verify webhook signatures and deduplicate event IDs.

Read the complete AgentMailer documentation for identity, email, A2A, events, and security semantics.