Plan by importance,
not just status —
and let your AI agents help.
AI is shipping work faster than any of us can hold in our heads. Project Tracker is the orientation layer for that new pace — plan by importance, keep your agents coordinated, and always see the wood for the trees.
Just enough tracker. Nothing you'll have to unlearn.
Two planning views, a day-at-a-glance Timeline, a shared backlog humans and agents can both drive — and an audit trail so nothing gets lost when work moves fast.
Value / Effort / Desirability matrix
Drag items into a 5×5 grid. Quadrants: Quick Wins, Big Bets, Fill-ins, Avoid. Reorder within cells by importance. Plan once, ship in priority order.
Kanban board for execution
Backlog → To Do → In Progress → Review → Done, with a Blocked column so stuck work stays visible. Drag-drop, subtask progress on parent cards, full-text search, milestone filters. The familiar bit.
Timeline — your day, at a glance
What you finished today, what's planned for tomorrow, and a rolling 14-day view of what actually got done. Drag work in and out as the day unfolds. Done-today picks up ad-hoc completions, not just things you formally scheduled.
AI-native via MCP
Your agents read the backlog, claim items, do the work, and append to the change log — through typed MCP tools or plain curl. Built for the way you actually work with Claude Code, Cursor, and anything else you reach for.
Multi-user, multi-AI — no clobbering
The board, matrix, and project bar update live as work happens — no refresh, no polling. When you and your agents are writing in parallel, edits never overwrite each other and nobody steps on a claimed item. You stop babysitting.
Audit trail with an undo
Stories for thinking, tasks for steps, linked. Append-only change log plus an activity timeline so you can always see who did what (human or AI). Deleted items go to a recoverable trash — mistakes are just an undo away.
Built for an AI-augmented backlog.
Most trackers were built for humans clicking buttons. Project Tracker was built knowing AI agents would be writing to it as much as people. Whether your agent speaks MCP, or is anything that can run
curl, it reads the backlog, claims items, does the work, and appends to the change log. Run several at once — the tracker stops them clobbering each other or working on the same thing twice.
get_item { key: "WEB-12" }
// → status: "todo", assignee: null
update_item {
key: "WEB-12",
status: "in_progress",
assignee: "claude-opus"
}
// ... agent does the work ...
update_item {
key: "WEB-12",
status: "review",
changeLog: "Added X to api/src/foo.ts; tests in api/__tests__/foo.test.ts"
}Opinionated about who this is for.
If you've bounced off heavier trackers, you're probably in the right place. If you need enterprise governance, you're not — and we'd rather say so up front.
For
YES- Founders, indie devs, and consultants juggling multiple projects
- Product, finance, and ops teams organising their own work
- Small teams (2–10) tired of Jira's ceremony
- First-time tracker users — a guided sign-up walks you in
Not for
NO- Enterprise sprint orchestration
- Regulated industries needing SSO/SAML
- Teams that want a deep dependency graph
- Anyone selling certifications
Currently invite-only.
Registration is open.
- 1 Register
At app.project-tracker.ai with your email and name.
- 2 Check your inbox
For a temporary password we'll email you.
- 3 Sign in
Choose a real password and start planning.