Active Development — DevTrack is being actively built. Features may change and some are still experimental. Follow progress on GitHub →
Local-First · No Cloud Required · Open Source

Stop managing.
Start shipping.

From a single problem statement, DevTrack plans your entire sprint — breaking it into Epics, Stories, and Tasks, assigning them based on team workload, and pushing everything to Azure DevOps, GitLab, or GitHub. Then it keeps every ticket in sync as you commit. All of it controllable from your phone.

Get Started Free Read the Docs
devtrack — zsh
100%
Local — your data stays on your machine
3+
PM integrations (Azure, GitLab, GitHub)
1
Sentence to a fully planned sprint
📱
Full control from your phone via Telegram

Built for developers
too busy to do admin work

DevTrack handles the gap between writing code and keeping your team in the loop — automatically.

Git-Aware Automation
Every commit you push is detected in real time. DevTrack extracts the ticket number, time, branch context, and PR info — then routes it through AI to update the right work item automatically. No copy-pasting.
fsnotify · Go daemon
🤖
git-sage Agent
A local LLM-powered git agent that plans and executes complex operations autonomously — squash, rebase, resolve conflicts — with full undo support.
Ollama · local LLM
Featured
🗂️
End-to-End PM Agent
Describe a problem in plain English and DevTrack plans the entire workstream — decomposing it into Epics, Stories, and Tasks, then assigning each item to the right person based on current workload. Works across Azure DevOps, GitLab, and GitHub. Trigger it from your desk or straight from your phone via Telegram.
plan · assign · create · sync · mobile
📱
Remote via Telegram
Full remote control from your phone. View work items, create tickets, trigger updates, read logs — all without opening your laptop.
Telegram bot
🎨
Talk Like You
Learns your communication style from Teams messages using RAG. Every AI-generated update sounds like you wrote it — not a bot.
ChromaDB · spaCy
🔌
Offline-First
Everything runs locally. DevTrack queues commits and updates when you're offline and syncs when connectivity returns. No cloud dependency.
SQLite · store-and-forward

From commit
to PM update
in seconds.

The entire pipeline runs locally with no cloud in the loop.

1

Git Commit Detected

fsnotify watches your repo. The moment you commit, the Go daemon fires a trigger with branch, PR, and diff context.

2

AI Enrichment

spaCy extracts entities, ticket numbers, and time estimates. The LLM enhances the message in your communication style.

3

PM Systems Updated

Matched work items get a comment, state transition, and time log. Missing items are created. All synced to Azure, GitLab, or GitHub.

4

You Get Notified

Telegram bot pushes a summary to your phone. Your standup writes itself.

devtrack pipeline
Powered by

Everything talks
to everything.

Native integrations with the tools you already use. No middleware, no SaaS, no subscriptions.

Azure DevOps
Work items, sprints, state transitions, comments
GitLab
Issues, milestones, assignment poller, webhooks
GitHub
PRs, issues, review requests, comments
Telegram
Remote control, notifications, /plan command
Ollama
Local LLM inference — no API keys needed
OpenAI / Groq
Cloud fallback when you need more power
MS Teams
Style learning from your chat history
SQLite
Local storage, offline queue, history

A day in the life
with DevTrack.

Smart Commit Messages
$ devtrack git commit -m "fix auth"
✦ Analyzing diff (23 lines, 2 files)...
✦ Detected: PR #142 · branch: fix/oauth-loop
✦ AI refining message...

✓ Suggested: "Fix OAuth redirect loop causing
infinite reauth cycles (#142)"

Accept [A] · Enhance [E] · Regen [R]: A

✓ Committed · Azure PR #142 commented
🤖 git-sage Agent
$ devtrack sage do "squash last 5 commits"
Mode: auto / review / suggest-only? auto

→ Plan: git reset --soft HEAD~5
git commit -m "..."

✓ Step 1/2: reset complete (HEAD~5 saved)
✓ Step 2/2: new commit created

Done. Any follow-up questions? (5 remaining)
🗂️ PM Agent via Telegram
You: /plan Build a payment processing system

Bot: Platform? [Azure] [GitLab] [GitHub]
You: Azure

✦ Decomposing with LLM...
✓ 1 Epic · 4 Stories · 12 Tasks

Preview ready. Create all 17 items?
You: Yes
✓ 17/17 items created in Azure DevOps
🎨 Personalized AI Updates
$ devtrack test-response "completed auth module"

✦ Loading style profile...
✦ Retrieving 3 RAG examples...

Generated (in your voice):
"Wrapped up the auth module — JWT
validation + refresh token rotation
both done. Ready for review."

Matches your comunication style ✓
👨‍💻
Shashank Raj
Software Engineer · Creator of DevTrack

I built DevTrack because I was spending too much time on the overhead that surrounds coding — updating tickets, writing standup notes, keeping PM tools in sync. The actual coding was the easy part.

DevTrack is my attempt at a developer-first automation layer: one that runs locally, respects your privacy, and handles the bureaucracy so you can stay in flow.

It also helps by offloading trivial work to low cost / no cost llm while allowing the frontier models to use up the precious tokens.

In fact, this project was planned and developed using Devtrack.

The vision
DevTrack should be like a senior engineer who never forgets to update the ticket — who knows what you're working on from your commits, writes the update in your voice, and keeps every stakeholder in the loop without you lifting a finger.
— The goal driving every feature decision
3+
PM integrations
100%
Local & private
MIT
Open source

Questions,
ideas, or
feedback?

DevTrack is open source and community-driven. Reach out through any of these channels.

Send a message

No tracking, no spam. Messages go straight to the maintainer.

Ready to stop doing
admin work manually?

Set up in 5 minutes. Runs locally. No accounts, no subscriptions, no privacy tradeoffs.