claude-tap: what it is, what problem it solves & why it's gaining traction

What it solves

claude-tap is a local proxy and trace viewer designed for AI coding agents. It allows developers to peek inside the "black box" of agentic workflows by capturing and visualizing the real API traffic between the agent and the LLM provider, revealing the exact context being sent and received.

How it works

The tool acts as a local proxy (supporting both reverse and forward proxy modes) that intercepts HTTP/SSE traffic from AI coding clients. It records the raw requests and responses, then provides a live browser-based dashboard to inspect these traces. It can also import local session transcripts from certain apps.

Who it’s for

Developers and AI researchers who use AI coding agents (like Claude Code, Cursor, or Codex) and need to debug agent behavior, inspect system prompts, or analyze token usage and tool calls.

Highlights

• Multi-Client Support: Works with a wide array of agents including Claude Code, Codex, Gemini CLI, Kimi, Cursor CLI, and others.
• Detailed Inspection: View system prompts, conversation history, tool schemas, tool calls, streaming responses, and token usage.
• Request Diffing: Compare adjacent requests to pinpoint exactly what changed in the prompt or context between turns.
• Privacy-Focused: Traces are kept on the local machine, and common authentication headers are redacted before recording.
• Portable Artifacts: Trace sessions can be exported as self-contained HTML files for easy sharing or archiving.