Sniffly provides a local dashboard for analyzing Claude Code usage, identifying error patterns, and sharing message histories. It targets developers and researchers working with Claude Code, offering insights to improve prompt engineering and collaboration.

How It Works

Sniffly processes local Claude Code log files, enabling users to visualize usage statistics, pinpoint common errors, and review past conversations. Its core advantage lies in providing a private, on-device analytics solution without sending data externally, ensuring user privacy.

Quick Start & Requirements

• Install via uv tool install sniffly@latest or pip install sniffly.
• Run initialization with sniffly init.
• Requires Python 3.10+.
• Access dashboard at http://localhost:8081.
• Official Documentation: https://sniffly.dev/

Highlighted Details

• Local processing ensures data privacy; no telemetry.
• Sharable dashboard feature with privacy options (private/public).
• Customizable settings for port, browser auto-open, and caching.
• Error breakdown and message history analysis for prompt improvement.

Maintenance & Community

The project is maintained by Chip Huyen. Further community engagement details are not explicitly provided in the README.

Licensing & Compatibility

Released under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The dashboard is designed for local execution; remote server deployment requires manual port forwarding. Configuration issues can be resolved via the CLI or by resetting the configuration file.