GitHub - Maciek-roboblog/Claude-Code-Usage-Monitor: Real-time Claude Code usage monitor with predictions and warnings
Real-time Claude Code usage monitor with predictions and warnings - Maciek-roboblog/Claude-Code-Usage-Monitor
Open Source Developer Tool
Summary
Sherlock is an open‑source Python tool that acts as a transparent proxy for LLM APIs, capturing request data and showing token consumption in a live terminal dashboard. It helps developers monitor costs, debug prompts, and keep track of context window usage across AI development sessions.
Highlights
- Intercepts LLM API traffic via a transparent CLI proxy
- Real‑time terminal dashboard with color‑coded fuel gauge for token usage
- Automatically saves prompts and responses as Markdown and JSON files
- Supports multiple providers including Anthropic Claude, OpenAI, and Google Gemini
- Easy installation with pip and simple commands like "sherlock start" or "sherlock claude"
auto-generated
Context
Audience
Developers and AI engineers who build applications with large language models and need to track token usage, debug prompts, and manage API costs
DomainLarge Language Model Development
Formatcommand-line application
Accessopen source on GitHub
Source Overview
Sherlock is a Python 3.10+ tool that intercepts LLM API traffic via a transparent proxy and displays real-time token usage in a terminal dashboard.
It tracks context window usage with a color-coded fuel gauge, saves prompts as Markdown and JSON files, and supports providers like Anthropic Claude Code, OpenAI Codex, and Google Gemini CLI. Installation uses git clone and pip install -e ., with commands like "sherlock start" for the dashboard and "sherlock claude" to proxy tools.
Topics
Related
LangChainOpenAI CLI toolsAnthropic Claude APIGoogle Gemini SDKLLM cost monitoring utilities
