pella-labs

pella-labs / bematist

Public

Open-source, self-hostable AI-engineering analytics platform

13
0
100% credibility
Found Apr 18, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Bematist is a self-hosted analytics platform that automatically tracks AI coding assistant usage on developers' machines and correlates it with code outcomes like commits and tests.

How It Works

1
🔍 Discover Bematist

You hear about Bematist from a colleague who uses it to understand how their team spends time with AI coding helpers.

2
💻 Install the tracker

Download and run a simple program on your computer that quietly watches your AI tools without getting in the way.

3
🏠 Set up your dashboard

Start a private web dashboard on your own computer or server with one easy command—it handles everything automatically.

4
📊 Watch insights appear

As you code, see real-time stats on AI usage, costs, and how it connects to your code changes and progress.

5
👥 Share with your team

Invite teammates to add the tracker so you can compare and learn together without sharing private details.

✅ Optimize your workflow

Use the clear reports to spend smarter on AI tools and boost your team's coding speed and quality.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is bematist?

Bematist is an open-source, self-hostable AI-engineering analytics platform built in TypeScript that auto-instruments developers' machines to track LLM and coding-agent usage—tokens, costs, prompts, sessions, tool calls—from tools like Claude Code, Cursor, Continue.dev, and Codex. It ships data to your own backend (Docker Compose with Postgres, ClickHouse, Redis) for a dashboard correlating AI spend with git outcomes like commits and PRs, surfacing insights such as why one dev uses half the tokens for similar work. Privacy-first: no per-engineer leaderboards or real-time surveillance, just aggregate views via opt-in tiers.

Why is it gaining traction?

As an open-source GitHub Copilot alternative for analytics—not coding—it stands out by observing without gating APIs or shipping leaderboards, dodging privacy pitfalls and TOS issues common in proprietary tools. Self-hostable like a Heroku/Netlify/Vercel alternative, it deploys via simple docker-compose up, with signed binaries for collectors and CLI commands like `bematist status` or `serve`. Devs hook it for honest ROI on AI tools without panopticon vibes.

Who should use this?

Engineering managers at AI-heavy teams tracking Cursor/Claude Code spend against git velocity. Self-hosting orgs wanting open-source GitHub Actions alternatives for telemetry pipelines. Early adopters in ai-engineering analytics experimenting with tenant-owned dashboards.

Verdict

Promising foundation for self-hostable AI analytics, with solid docs, privacy gates, and tests—but 13 stars and 1.0% credibility score signal pre-1.0 maturity; expect rough edges. Try if you're building bematist workflows now; watch for M1 polish.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.