Set up Cursor

Create an Agiler API key with projects:read, restricted to the project you want to inspect.

Provide the key

Make AGILER_API_KEY available in the environment that launches Cursor. For a POSIX shell with the cursor command installed, fully quit Cursor, then run:

export AGILER_API_KEY='ak_REPLACE_WITH_YOUR_KEY'
cursor .

Replace the placeholder locally. For regular use, supply the key through your normal secret manager or private environment configuration. Shell history may retain the export command.

Configure the server

Create .cursor/mcp.json in your project, or use ~/.cursor/mcp.json for a personal configuration across projects. Merge this entry into any existing mcpServers object:

{
  "mcpServers": {
    "agiler": {
      "url": "https://mcp.agiler.io/v1",
      "headers": {
        "Authorization": "Bearer ${env:AGILER_API_KEY}"
      }
    }
  }
}

Cursor uses ${env:AGILER_API_KEY} for environment substitution. Keep that reference in the configuration so each user can provide their own secret. Remote HTTP servers do not use Cursor’s envFile setting. See Cursor’s MCP documentation for configuration details.

Verify in Agent

Restart Cursor after saving the configuration. Open Customize → MCPs and check that agiler is enabled. In an Agent conversation, ask:

Use Agiler’s whoami tool to check my account, then list the projects I can access. Do not make changes.

Review any tool-approval prompts and check the results. Continue with your first task or add project instructions.

If tools are unavailable

Open the Output panel and select MCP Logs to inspect connection errors. An authentication failure often means the Cursor process did not receive the environment variable. Restart it from an environment containing the key. Organization policies can also restrict servers or tools. See Cursor’s MCP help and Agiler troubleshooting.