ccglass
ccglass: A Lightweight Local Reverse-Proxy for AI Coding Agents
Overview
ccglass is a lightweight, local reverse-proxy and web dashboard designed to provide deep visibility into the traffic of AI coding agents. It allows developers to see exactly what their coding agents send to models in real-time. It supports a wide range of tools including Claude Code, Codex, DeepSeek-TUI, Reasonix, Kimi, OpenCode, Ollama, OpenRouter, and more.
Benefits
ccglass solves a common problem for developers. Many AI coding agents bypass standard HTTP proxy settings, making it hard to monitor their traffic with traditional tools. ccglass sidesteps these issues by acting as a local reverse-proxy. This approach involves a simple plain HTTP hop where the agent sends requests to the local proxy. The proxy then handles the secure HTTPS connection to the real API. This method means there is no need to install CA certificates or manage complex TLS pinning in the agent's environment. The tool is easy to install using npm and requires no runtime dependencies for the core proxy. It works with major AI coding agents and model providers via specific commands that set the appropriate environment variables. The dashboard provides a comprehensive view of the agent's interaction with the model. Users can see live request streams, conversation flow diagrams, and token analysis. It also shows exact input and output tokens, cache-hit rates, and estimated costs per request. The tool allows users to compare requests to see what context was added in specific turns. It displays per-request latency, including total time and time to first token. Users can export requests as raw HTTP transcripts, Markdown, JSON, or HAR files. The dashboard includes session summaries with rolled-up statistics for the full session. It supports both light and dark modes with per-model filters. When wrapping Claude Code, ccglass exposes itself as an MCP server, allowing the agent to inspect live traffic directly from within a session.
Use Cases
ccglass is useful for developers who want to understand how their AI coding agents work. It is ideal for teams that need to audit AI interactions for security or compliance reasons. Developers can use it to analyze token usage and control costs by seeing exactly how many tokens are consumed. It helps in debugging issues by allowing users to see the exact prompts and responses in real-time. The tool is compatible with IDE extensions that allow configuring a custom API base URL, such as Cursor, Cline, and Continue.dev. It works with OpenAI-compatible IDEs by running a proxy command. It also works with Anthropic-compatible IDEs by running a proxy command. This setup works in Bring Your Own Key or custom-API-key modes. It cannot intercept built-in subscription routes through specific API endpoints. The tool is suitable for anyone using AI coding agents who wants a transparent method for inspecting traffic without the friction of traditional proxy setups.
Pricing
Pricing details are not available in the provided information.
Vibes
Pricing details are not available in the provided information.
Additional Information
Pricing details are not available in the provided information.
This content is either user submitted or generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral), based on automated research and analysis of public data sources from search engines like DuckDuckGo, Google Search, and SearXNG, and directly from the tool's own website and with minimal to no human editing/review. THEJO AI is not affiliated with or endorsed by the AI tools or services mentioned. This is provided for informational and reference purposes only, is not an endorsement or official advice, and may contain inaccuracies or biases. Please verify details with original sources.
Comments
Please log in to post a comment.