Attio Model Context Protocol (MCP) server implementation
Configuration
View docs{
"mcpServers": {
"fvefwfe-attio-mcp-server": {
"command": "attio-mcp-server",
"args": [],
"env": {
"ATTIO_API_KEY": "YOUR_API_KEY",
"ATTIO_WORKSPACE_ID": "YOUR_WORKSPACE_ID",
"ATTIO_DEFAULT_CURRENCY": "USD",
"ATTIO_DEFAULT_DEAL_OWNER": "[email protected]",
"ATTIO_DEFAULT_DEAL_STAGE": "Interested"
}
}
}
}The Attio MCP Server lets you talk to your Attio CRM through natural language using AI assistants like Claude. It translates your everyday language into actions on Companies, People, Lists, Tasks, and other records, so you can find, update, and analyze data without navigating menus or writing API calls.
You interact with the MCP server through an MCP client (such as Claude Desktop) by asking natural language questions or issuing commands. The server translates your requests into universal tools that cover searching, creating, updating, and organizing records across Companies, People, Lists, and Tasks. You can perform complex queries with advanced filters, move deals through pipelines, manage batch operations, and export data. Start with simple requests like finding records or adding a task, then evolve to multi-step actions that combine several operations in one flow.
Prerequisites you need before you install the MCP server are: Node.js version 18 or higher and a valid Attio API Key for your workspace. You also need your Attio Workspace ID. Once you have these, choose one of the installation methods below.
Configure field mappings to translate natural language into Attio API field names. Create your user mapping file by copying the sample and tailoring it to your workspace's custom fields under the objects.companies and objects.people sections. The mappings are merged with the defaults so only your overrides and additions are stored.
Here is a representative Claude Desktop MCP configuration snippet showing how to wire the server and set environment variables (values shown as placeholders):
{
"mcpServers": {
"attio-mcp": {
"command": "attio-mcp-server",
"env": {
"ATTIO_API_KEY": "your_api_key_here",
"ATTIO_WORKSPACE_ID": "your_workspace_id_here",
"ATTIO_DEFAULT_DEAL_STAGE": "Interested",
"ATTIO_DEFAULT_DEAL_OWNER": "[email protected]",
"ATTIO_DEFAULT_CURRENCY": "USD"
}
}
}
}The MCP server authenticates requests with API keys and operates with direct Attio API access, with no local data retention on the server. You can deploy on your own infrastructure if you prefer on-premises. If you encounter issues, refer to the troubleshooting guidance for common problems like field mapping mismatches, pagination quirks, or missing endpoints.
Common issues include ensuring the field mappings cover your custom fields and handling pagination properly for large result sets. Review your environment variables and the field mapping overrides to ensure smooth operation.
Sales teams can search for high-potential prospects, update records in bulk, and create tasks for follow-ups in one natural-language session. Marketing teams can pull engagement metrics, segment lists, and export data for campaigns. Customer success can monitor renewals, track activity history, and generate reminders for check-ins.
If you run into issues, check the latest guidance on troubleshooting and known limitations, such as certain filter capabilities or pagination behaviors. There are documented fixes and improvements for common problems, along with status updates on API compatibility and performance optimizations.
Perform broad or targeted searches across Companies, People, Lists, and Records using unified search tools.
Build complex multi-criteria queries with AND/OR logic and timeframe capabilities.
Create, update, or delete multiple records in chunks with error handling.
Universal create, read, update, and delete operations across resource types.
Find connections between entities, such as linking people to companies or lists to records.