About Sniffly
Sniffly is a dashboard tool built specifically for people using Claude Code, Anthropic's coding agent CLI, who want to understand their own usage patterns in more depth. It analyzes the logs Claude Code already generates locally to surface insights that help a developer use the tool more effectively.
Its main features are understanding your own usage patterns over time, breaking down where Claude Code made mistakes so similar issues can be avoided in the future, and walking through full message history, which can optionally be shared with coworkers. A built-in sharing feature lets a user generate a link to their dashboard with configurable privacy, choosing between a private link, a public listing in a gallery, or whether to include the actual command text in what is shared.
Sniffly runs locally as a web server, by default accessible at localhost:8081, and can be installed with uv, pip, or run directly from source; configuration options cover the port, host, caching limits and how many messages load initially. It requires Python 3.10 or newer, is written in Python, and is released under the MIT license.
Key features
- Claude Code usage pattern analytics
- Error breakdown and mistake analysis
- Full message history browsing
- Shareable dashboard links with privacy controls
- Local web server, no data leaves your machine by default
Good fit for
- Understanding your own Claude Code usage habits
- Sharing coding agent instructions and stats with a team
Sniffly: questions and answers
- What is Sniffly used for?
- Sniffly is a free, MIT-licensed, self-hosted analytics dashboard for Claude Code that analyzes local usage logs to surface usage patterns, error breakdowns and shareable reports. It is a good fit for understanding your own Claude Code usage habits, and sharing coding agent instructions and stats with a team.
- Is Sniffly open source?
- Yes. Sniffly is open source under the MIT licence. Its source code is on GitHub at chiphuyen/sniffly and is written mainly in Python.
- Is Sniffly free?
- Yes. Sniffly is open source, so the software itself is free to use.
- Can I self-host Sniffly?
- Yes. Sniffly can be self-hosted on your own server or infrastructure; there is no official hosted version.
- What are some alternatives to Sniffly?
- Similar open-source tools in the Analytics category include Litlyx, Umami and Ackee. SaaS products in the same category include Toucan Toco, AffiliateTrack and Analyse.
- Is Sniffly actively maintained?
- The most recent commit to Sniffly was on 8 August 2025. The project has 1.3k stars on GitHub, with 10 added in the last 30 days.
Open-source alternatives to Sniffly
See all
Litlyx
Analytics
Powerful Analytics Solution. Setup in 30 seconds. Display all your data on a Simple, AI-po
Apache-2.0vs Google Analytics★ 1.7k
Umami
Analytics
Umami is a privacy-first analytics platform. Traffic, campaigns, behavior, conversions, an
MITvs Google Analytics★ 40k
Ackee
Analytics
Self-hosted, Node.js based analytics tool for those who care about privacy.
MITvs Google Analytics★ 4.7k
Aptabase
Analytics
✨ Open Source, Privacy-First and Simple Analytics for Mobile, Desktop and Web Apps
AGPL-3.0vs Google Analytics★ 1.9k
Tianji
Analytics
Tianji: Insight into everything, Website Analytics + Uptime Monitor + Server Status. not o
Apache-2.0vs Google Analytics★ 3.1k
PostHog
Analytics
PostHog automatically diagnoses problems, fixes bugs, and generates pull requests – all wi
OSSvs Mixpanel★ 41k
SaaS alternatives to Sniffly
See all
Toucan Toco
Analytics
AI-powered custom analytics platform for building dashboards and data apps
SaaS
AffiliateTrack
Analytics
Dashboard that combines web analytics and affiliate network data for affiliate publishers
SaaS
Analyse
Analytics
Web analytics with funnels and cohorts, an AI SEO content engine and a data copilot
SaaS
Flowsery
Analytics
Analytics platform built around real user journeys
SaaS
Launchly
Analytics
iOS app with Home and Lock Screen widgets for tracking Stripe customer lifetime value
SaaS
MRR Empire
Analytics
Single dashboard that brings all your Stripe accounts together
SaaS

