Debug LLM Traces From Your IDE Using Natural Language MCP Queries

The MCP server lives at https://api.futureagi.com/mcp and uses OAuth. No API keys to copy around.

Claude Code Cursor VS Code Claude Desktop Windsurf

claude mcp add futureagi --transport http https://api.futureagi.com/mcp

After running, claude mcp list should show futureagi with ! Needs authentication. That’s expected; the OAuth handshake happens in the next step.

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "futureagi": {
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Or use the one-click install link on the setup page.

Add to .vscode/settings.json:

{
  "mcp.servers": {
    "futureagi": {
      "type": "http",
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "futureagi": {
      "url": "https://api.futureagi.com/mcp"
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "futureagi": {
      "serverUrl": "https://api.futureagi.com/mcp"
    }
  }
}

Restart your IDE after editing the config.

The first MCP tool call opens a browser to the consent screen. Review the 14 permission groups, click Authorize. Token cached, done.

Open your IDE’s chat panel and ask. The MCP server exposes ~50 trace-related tools (search, error analysis, span trees, error clusters), so phrase questions naturally. Your assistant picks the right tools.

Find failing traces:

List the most recent traces in my project that have errors.

Calls search_traces with has_error=True. If no traces have raw error flags, your assistant pivots to list_error_clusters and surfaces the AI-detected error categories across your projects. A richer signal than HTTP errors alone.

Inspect a specific trace:

Show me the span tree for the second trace from the previous list.

Calls get_span_tree. Returns the parent span plus nested LLM/tool calls with timing and inputs.

Diagnose what went wrong:

Run error analysis on that trace.

Calls get_trace_error_analysis. Returns categorized findings (hallucination, wrong intent, tool misuse) with severity and a quality scorecard.

Look across the project:

Analyze all traces in my project from the last hour and group failures by category.

Calls analyze_project_traces and list_error_clusters. Returns a histogram with the dominant error types.

Score or annotate from chat:

Add the tag needs-policy-grounding to the failing traces, and annotate them with “fabricated specifics, needs RAG over policy docs.”

Calls add_trace_tags + create_trace_annotation per matching trace. The annotations show up in the dashboard immediately.

The same chat that read the trace can now read your code. Ask:

Based on the error analysis, draft a system-prompt patch that refuses to answer policy questions when no grounding tool is available. Show it as a diff against agent.py.

Your assistant has both the trace findings (from MCP) and the file (from your editor). It produces a paste-ready diff. Apply it, re-run a few queries through the agent, and ask the next turn:

Re-check the latest traces in my project and confirm the fabrication category dropped.

That’s the full loop. Failure detection, diagnosis, fix, verification, all driven from one IDE chat thread.