# liaohch3/claude-tap

Intercept and inspect Coding Agent API traffic from Claude Code, Codex CLI, Gemini CLI, Cursor CLI, OpenCode, Kimi/Kimi Code, Pi, and Hermes in a local trace viewer.

Repository: https://github.com/liaohch3/claude-tap
Canonical: https://ross.abutalabs.com/products/claude-tap
Language: Python
License: MIT
License Family: permissive
Topics: ai-tools, claude-code, codex, developer-tools, llm, proxy, trace, agent-debugging, agent-observability, ai-agents, api-debugging, codex-cli, cursor-cli, gemini-cli, hermes-agent, kimi, llm-proxy, opencode, pi-coding-agent, trace-viewer
Last push: 2026-08-26T03:05:32+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 86, longevity 14
- inputs: {"age_days": 199, "days_push": 7, "days_rel": 17, "gap_med": 0.0, "n_releases_24m": 125}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3125, forks 269 (observed 2026-08-28T04:07:44.442025+00:00)

## What it is
claude-tap is a local proxy and trace viewer for AI coding agents like Claude Code, Codex CLI, and Gemini CLI. It intercepts real API traffic so you can inspect system prompts, conversation history, tool schemas, streaming responses, token usage, and request diffs in a local viewer.

## Use cases
- inspect what system prompts my coding agent actually sends
- debug Claude Code API requests locally
- view token usage and tool calls from Codex CLI
- compare agent context between turns with request diffs
- trace LLM API traffic from Gemini CLI or Cursor CLI
- understand why my AI agent behaves unexpectedly
- capture and review agent conversation history locally

## When to choose
- you use AI coding CLIs and want full visibility into their API traffic
- you need to debug agent context, prompts, or tool schemas
- you want a local, privacy-preserving trace viewer without sending data anywhere

## When to avoid
- you need production LLM observability for your own app's backend
- you want hosted team-wide tracing dashboards
- you don't use AI coding agents at all

## Facets
- artifact type: cli-tool
- maturity: active
- function: proxy, monitoring, tracing, developer-tools, llm-inference
- domain: developer-tools, large-language-models, apis
- platform: cli, python, cross-platform, windows
- tags: llm-proxy, trace-viewer, agent-debugging, api-debugging, claude-code, codex, gemini-cli, cursor-cli, context-engineering, agent-observability, ai-agents, command-line, macos, linux

## Member repositories
- liaohch3/claude-tap (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:44.442025+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:26:23.491664+00:00, confidence not recorded.
  - readme: https://github.com/liaohch3/claude-tap (fetched 2026-08-28T04:07:44.442025+00:00, sha 3de7aa11b900)
  - registry_pypi: https://pypi.org/pypi/claude-tap/json (fetched 2026-08-29T09:41:20.553296+00:00, sha 9e506f1b7395)
- Data as of 2026-08-30T08:39:29.467469+00:00.