Cursor reads the exception you just hit

1. Turn MCP on

MCP is off by default. Start Buggregator with this variable, or set mcp.enabled: true in buggregator.yaml.

MCP_ENABLED=true

2. Add the server to mcp.json

Cursor reads MCP servers from .cursor/mcp.json in your project, or from ~/.cursor/mcp.json for every project. The command and its argument go in separate fields. buggregator is the binary: the name works when it is on your PATH, otherwise give the full path.

{
  "mcpServers": {
    "buggregator": {
      "command": "buggregator",
      "args": ["mcp"]
    }
  }
}

3. Use the HTTP URL if Buggregator runs in Docker

Run Buggregator with the HTTP transport, then give Cursor a url instead of command and args.

services:
  buggregator:
    image: ghcr.io/buggregator/server:latest
    ports:
      - 127.0.0.1:8000:8000
      - 127.0.0.1:8001:8001    # MCP HTTP port
    environment:
      MCP_ENABLED: "true"
      MCP_TRANSPORT: http
      MCP_ADDR: ":8001"

{
  "mcpServers": {
    "buggregator": {
      "url": "http://localhost:8001/mcp"
    }
  }
}

4. Ask about the error

Trigger the error, then ask Cursor about it. The assistant picks the tools.

"Show me the latest Sentry errors"
"What's the stack trace for this exception?"
"What value was dumped in the last VarDumper event?"

Buggregator has a built-in MCP server, and mcp.json is Cursor's own format (https://cursor.com/docs/mcp). sentry_event returns an error's message, severity, exception chain with stack traces and environment; vardump_get returns a dump as plain text; profiler_summary and profiler_top give the slowest functions. Cursor reads the data your app already sent, so nothing is copied by hand.