Analyze Claude Code logs to help you use Claude Code better.
See where Claude Code makes mistakes so that you avoid these mistakes.
Walk through all your instructions and share them with your coworkers if needed.
- Requirement: Python 3.10+
Make sure you have uv installed! https://github.com/astral-sh/uv
# One-time execution (no installation needed)
uvx sniffly@latest init
# Install the package
uv tool install sniffly@latest
sniffly init
After running sniffly init, access your dashboard at http://localhost:8081 (or whichever port you choose).
pip install sniffly
sniffly init
git clone https://github.com/chiphuyen/sniffly.git
cd sniffly
pip install -e .
sniffly init
# Change port (default: 8081)
sniffly config set port 8090
# Disable auto-opening browser
sniffly config set auto_browser false
# Show current configuration
sniffly config show
port | 8081 | Server port |
host | 127.0.0.1 | Server host |
auto_browser | true | Auto-open browser on start |
cache_max_projects | 5 | Max projects in memory cache |
cache_max_mb_per_project | 500 | Max MB per project |
messages_initial_load | 500 | Initial messages to load |
max_date_range_days | 30 | Max days for date range selection |
See full CLI Reference for all options and commands.
You can create a link to share your project's stats and instructions with your coworkers.
- Click the "📤 Share" button in your dashboard
- Choose privacy options:
- Private: Only people with the link can view
- Public: Listed in the public gallery
- Include Commands: Share your actual command text
- Copy and share the generated link
Port already in use?
# Use a different port
sniffly init --port 8090
# Or change default
sniffly config set port 8090
Browser doesn't open?
# Check setting
sniffly config show
# Enable auto-browser
sniffly config set auto_browser true
# Or manually visit http://localhost:8081
Configuration issues?
# View all settings and their sources
sniffly config show
# Reset a setting to default
sniffly config unset port
# Remove all custom config
rm ~/.sniffly/config.json
For more issues, see GitHub Issues.
Sniffly runs entirely on your local machine:
- ✅ All data processing happens locally
- ✅ No telemetry
- ✅ Your conversations never leave your computer
- ✅ Shared dashboards are opt-in only
MIT License - see LICENSE file.
- Homepage: sniffly.dev
- Documentation: Full CLI Reference
- Issues: GitHub Issues