#19 · Primary category: AI Usage & Cost Monitoring
ccglass
See what your coding agent (Claude Code, Codex, Kimi) sends to the model — local proxy + web dashboard
Project last updated:07/09/26
GitHub Stars
764
Forks
48
Contributors
10
License
MIT
Why we included this project
Most coding agents ignore the HTTP_PROXY and HTTPS_PROXY environment variables, which is why the usual Charles or mitmproxy tricks never see their traffic. ccglass works around that by running a local proxy that the agent talks to over plain HTTP on localhost, while the agent still does the HTTPS to the real API itself. That means no CA certificates to install and no TLS pinning to fight. Point it at Claude Code, Codex, or any of the other supported clients, and a web dashboard opens showing every request in real time: the full system prompt, every tool schema, message history, token and cache counts, estimated cost, and a turn-to-turn diff that makes prompt inflation easy to spot. It also works with IDE extensions like Cursor BYOK and Cline, and can expose itself as an MCP server so Claude can inspect its own traffic mid-session.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
CodexBar
Show usage stats for OpenAI Codex and Claude Code, without having to login.
codeburn
Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Codex, Gemini and more), by model, project, and task. npx codeburn
Claude-Code-Usage-Monitor
Real-time Claude Code usage monitor with predictions and warnings
helicone
🧊 Open source LLM observability platform. One line of code to monitor, evaluate, and experiment. YC W23 🍓
agentsview
Local-first session search, analytics, insights, and token use statistics for coding agents, supporting Claude Code, Codex, and more than 20 other agents.