Documentation Index
Fetch the complete documentation index at: https://docs.tight.com/llms.txt
Use this file to discover all available pages before exploring further.
The Tight Model Context Protocol (MCP) server provides a set of tools that AI agents can use to interact with the Tight
API. You can use the MCP tools as a developer embedding Tight into your product, as a business owner looking for
financial insights, or as your platform seeking to perform comparative analytics across all businesses on your platform.
Connect to Tight’s MCP Server
Claude Code
ChatGPT
Cursor
VS Code
Custom Connectors
Execute the following command to register the Tight MCP server. To learn more, see the Claude Code
documentation.claude mcp add-json tight '{"type":"http","url":"https://mcp.sandbox.tight.com/","headers":{"Authorization":"Bearer ${tight_access_token}"}}'
For ChatGPT Pro, Plus, Business, Enterprise, or Education users, follow the
OpenAI documentation for custom connectors using the
following parameters:
- Server URL:
https://mcp.tight.com
- Connection mechanism: Mixed Authentication
Add the following to you ~/.cursor/mcp.json. To learn more, see the Cursor
documentation.{
"mcpServers": {
"tight": {
"url": "https://mcp.tight.com",
"headers": {
"Authorization": "Bearer ${env:tight_access_token}"
}
}
}
}
Add the following to your .vscode/mcp.json. To learn more, see the VS
Code documentation.{
"servers": {
"tight": {
"type": "http",
"url": "https://mcp.tight.com",
"headers": {
"Authorization": "Bearer ${input:tight_access_token}"
}
}
}
}
Use the remote server URL https://mcp.tight.com with an accessToken obtained via the
Tight Authentication Endpoint.{
"mcpServers": {
"tight": {
"command": "npx",
"args": [
"mcp-remote",
"https://mcp.sandbox.tight.com",
"--header",
"Authorization:${AUTH_HEADER}"
],
"env": {
"AUTH_HEADER": "Bearer ${tight_access_token}"
}
}
}
}
For products connecting their existing AI chat interfaces to Tight’s MCP server, use the Custom Connectors. 
