Cursor + Diaphora

Connect Diaphora as an MCP server inside Cursor so the agent can run structured workflows without leaving the editor.

MCP Server URL

https://mcp.diaphora.ai/mcp

Commands & settings

mcp.json entry

Add this under mcpServers in ~/.cursor/mcp.json (global) or .cursor/mcp.json (project-only).

Overview

How Diaphora fits into Cursor

Cursor's agent can call Diaphora directly through MCP. Diaphora handles connector access, structured extraction, and repeatable plan execution; Cursor's agent treats the results as grounded context while you keep working in the editor.

Developers lean on Cursor's agent for day-to-day editing, but connector access, repeatable extraction, and schema-shaped output are still best handled by Diaphora plans. Adding Diaphora as an MCP server lets the agent call those plans directly from chat.

The agent stays focused on reading and writing code, while Diaphora handles the deterministic parts of the workflow such as gathering data from connected tools and returning structured, reliable results the agent can act on.

Setup

How to connect Diaphora to Cursor

Diaphora is exposed as a remote MCP server at https://mcp.diaphora.ai/mcp. Cursor connects to it through an mcp.json entry — add it once, authenticate, and the tools show up in every chat.

1. Create your Diaphora account
   Cursor authenticates through your Diaphora identity, so get an account set up first if you don't already have one.

2. Add Diaphora to mcp.json
   Open Cursor Settings → Tools & MCP → New MCP Server (this opens mcp.json), or edit the file directly at ~/.cursor/mcp.json for every project or .cursor/mcp.json for just this one. Add a "diaphora" entry with the URL above — see the reference block below for the exact JSON.

3. Sign in through Diaphora IAM
   Back in Settings → Tools & MCP, Diaphora shows a Needs login status the first time. Click it and Cursor hands off to Diaphora's IAM login in your browser — authorize there and Cursor picks the session back up automatically.

4. Start calling Diaphora plans from chat
   Once authorized, Diaphora's tools sit alongside Cursor's other MCP servers and the agent can call them mid-conversation — no need to reauthenticate until the token expires.

How it works

Working with the integration

1. Run the deterministic workflow through Diaphora
   Reach for a Diaphora plan when the task needs connector access, repeatable extraction, or schema-shaped output rather than free-form generation.

• The agent calls the plan as an MCP tool instead of trying to reconstruct the workflow from scratch.
  • Diaphora enforces structure before any result reaches the conversation.

2. Let the agent use the structured result as context
   The MCP tool call returns a structured, schema-shaped result that the agent can read, summarize, or use to inform the next edit.

• This cuts down on hallucinated details, since the agent is grounded in real tool output.
  • Results stay consistent across runs because the same plan enforces the same structure.

3. Ask the agent to act on the result, not re-derive it
   Use Cursor's agent for what it does best — reading, writing, and refactoring code — based on the facts Diaphora already gathered.

• Ask for code changes, tests, or documentation grounded in the plan's output.
  • Keep the factual backbone anchored to the Diaphora result rather than the model's own recall.

4. Rerun the plan when the facts change
   When the underlying data changes, rerun the plan through MCP instead of asking the agent to regenerate context by hand.

• Results stay consistent because the plan, not the agent's memory, is the source of truth.
  • The agent spends its time on code, not on reconstructing context.