AI integrations and MCP

Connect your AI assistant to the official documentation to build SendEasy integrations using documented endpoints, payloads and authentication.

Choose an access method

Method	Purpose	Address
`llms.txt`	Discover relevant documentation	English index
`llms-full.txt`	Read all English documentation at once	Full English documentation
Markdown	Read a specific page without interface HTML	Public API example
MCP	Let the assistant search and read references on demand	`https://docs.sendeasy.app/mcp`

Portuguese indexes are at /llms.txt and /llms-full.txt; Spanish indexes are at /es/llms.txt and /es/llms-full.txt. Append .md to a documentation page URL to read its Markdown version. Home pages use /index.md, /en/index.md and /es/index.md.

Connect with MCP

Add a remote MCP server in a client that supports Streamable HTTP:

Setting	Value
Name	`sendeasy-docs`
URL	`https://docs.sendeasy.app/mcp`
Transport	Streamable HTTP
Authentication	None

Configuration formats vary by client. For clients accepting mcpServers entries with a url:

{
  "mcpServers": {
    "sendeasy-docs": {
      "url": "https://docs.sendeasy.app/mcp"
    }
  }
}

Available tools:

Tool	Usage
`search_docs`	Search with `query`, `locale` (`pt`, `en`, `es`) and `limit` (1 to 10).
`get_doc`	Read the exact `path` returned by search. For long pages, pass the returned `nextOffset` as `offset` to continue.
`list_docs`	List pages and descriptions in the requested `locale`.

The default language is Portuguese. Set locale to en for English search results. get_doc follows the path language, for example /en/server/webhooks. MCP resources also expose indexes at sendeasy://docs/pt, sendeasy://docs/en and sendeasy://docs/es.

This MCP server reads public documentation. It does not send messages, modify contacts or access accounts. Do not provide customer tokens to the documentation server. Your integration runs in your own system through the SendEasy API using the appropriate credentials.

Example prompt

Use the official SendEasy documentation to build a Node.js integration
that sends a WhatsApp message when a lead submits my form.

Search for /api/messages/send and read the full reference in English.
Confirm the token type, required fields and phone number format.
Use an environment variable for the token, handle errors, and distinguish
queue acceptance from final delivery. Cite your sources and do not invent fields.

Without MCP, ask the assistant to start at https://docs.sendeasy.app/en/llms.txt and follow the relevant Markdown links.

API authentication

The API base URL is https://backend.sendeasy.app. Use Authorization: Bearer YOUR_TOKEN with the token appropriate to the endpoint:

Operation	Token	Reference
`/api/messages/send`	Channel token	Public API
`/api/v1/waba/*`	API Token	WABA API
`/api/v1/email/*`	Email channel token	Email API
`/api/v1/ai/*`	API Token	Bots API
`/api/integration/*`	Integration token	Custom integration

Keep credentials on your application server. Read the Technical Reference and webhook formats. An accepted or queued request does not confirm final message delivery.