Show HN: Understand and reduce token usage with ContextSpy context profiler (github.com)

ContextSpy, a new context window profiler designed for large language models (LLMs) and AI coding tools, has been introduced to help developers understand and optimize token usage. By intercepting requests to LLM APIs, ContextSpy provides a detailed analysis and visualization of prompt composition while tracking context changes throughout sessions. This is particularly significant as the costs associated with AI agent token consumption are rising, especially with workflows generating a disproportionate number of input tokens compared to outputs. By making token consumption patterns visible, developers can better manage their API expenses and performance. The tool operates as an HTTPS proxy, capturing and analyzing each request locally, which ensures user data remains private. ContextSpy enables users to monitor token categories such as system prompts, tool definitions, and conversation history in real-time, helping to identify inefficiencies. With a setup compatible with macOS, Linux, and Windows, it promises broad accessibility, and its features include session tracking and detailed dashboards. By highlighting the costly implications of excessive context and allowing for fine-tuning of input data, ContextSpy emerges as an essential resource for developers looking to maximize the efficiency of their AI applications.