HashMeterAi is an open-source solution for tracking real-time token consumption across various AI coding tools, such as Claude Code, Codex, Kimi, and Qwen CLI, while ensuring complete data privacy.
What Happened
HashMeterAi has been introduced—a local tool that works entirely offline by analyzing local transcript logs. The application provides a unified dashboard displaying computation costs in dollars, the number of processed tokens, and overall activity statistics. The tool does not have access to chat content or user secrets.
Context
When using multiple AI agents and CLI tools, developers often face the problem of opaque expenses and a lack of ability to promptly control budgets. Existing monitoring methods often require transferring data to third-party servers, which creates security risks for code and prompts.
Why It Matters for the Industry
This project opens a niche for "Local-first Observability" for AI agents. This creates demand for cost observability infrastructure tools that allow for transparent auditing of resource usage on the client side without sacrificing data privacy.
Why It Matters for Users
Developers gain a convenient way to visualize and aggregate token costs for local development tools. This allows for precise budget control and real-time optimization of AI assistant usage while maintaining full confidence in data security.
Sources
Author
Look at AI, Editorial Team