All work
Our productOur product · SaaS issue tracking

Kevta

Kevta is Matrix Minds’ own product: an issue tracker for software teams of two to fifteen who ship on GitHub. We took it from product idea to a running, billable beta — design system, API, realtime app, payments and all.

kevta.appVisit the live site
Kevta landing page — one issue tracker for the whole team, with a board preview
Type
Our own product
Stage
Beta
Scope
Product, design system, web app, API, billing
Platform
Web

The challenge

Small teams pick an issue tracker and then spend a week configuring it. Schemes, sprints, permission matrices and per-seat billing all arrive before the first bug is filed — and per-seat pricing quietly punishes teams for inviting a contractor or a founder who checks in on Fridays.

We wanted to find out whether a tracker could be opinionated enough to be useful on the first day, and priced so that adding a person is never a budget conversation.

The solution

Every board arrives with Title, Assignees, Status, Priority and Due Date already on it, and four views — board, list, table and calendar — that switch without losing context. There is nothing to configure before filing the first issue.

Billing is flat per workspace rather than per seat, which turned into a real engineering constraint: plan limits (workspaces, boards, members) have to be enforced server-side on every write, and Stripe has to stay in sync with them.

The Git integration is the part we care most about. Copy a ready-made branch name from any issue; when the pull request opens, Kevta matches it back by issue key with no linking step, and an automation rule can move the issue to Done the moment it merges.

Key capabilities

Boards, issues and four views

Board, list, table and calendar over the same data, with grouping, sorting and filtering per view.

A personalised home

Issues assigned to you, boards you pinned, recently-viewed boards and the activity you missed — instead of a generic task list.

GitHub integration

Branch-name generation, pull-request matching by issue key, and automation rules that move issues as checks pass and merges land.

Realtime collaboration

Socket.IO-backed live updates and notifications, so a board stays current without a refresh.

Workspaces and roles

Workspaces, roles and per-board membership keep the right people in the loop and everyone else out of the noise.

Subscriptions and plan limits

Stripe-managed subscriptions with plan limits enforced in the API — including receipts, tax and cancellation handling.

An MCP server

Kevta exposes a Model Context Protocol endpoint, so AI agents can read and update issues through the same authorised API as the UI.

Observability

Structured logging, Sentry error tracking and a documented API — the parts that make a product supportable rather than demoable.

Screens

Kevta pricing page showing the Free and Pro plans with workspace, board and member limits
Flat per-workspace pricing — the plan limits are the product boundary, so they are stated on one page.
Kevta documentation page explaining how to manage Kevta from an AI assistant over MCP
The MCP endpoint documented for users: an AI client acts on your boards through the same permissions and API as the UI.
Kevta landing page on a phone-sized viewport
The same layout system down to a 393px phone viewport.

Technology

  • Next.js (App Router)
  • React
  • TypeScript
  • Redux Toolkit
  • CSS Modules
  • Express 5
  • Prisma
  • PostgreSQL
  • Redis
  • Socket.IO
  • Stripe
  • Google OAuth
  • Model Context Protocol
  • Sentry

Confirmed from the project’s own codebase.

What we built

  • A production Next.js application and an Express + Prisma + PostgreSQL API, running with Redis, Socket.IO, Stripe subscriptions, Google OAuth, structured logging and Sentry.
  • A full design-system pass across the app: one component set, one type and colour system, a persistent app shell, and a rebuilt Home and Settings information architecture.
  • A data-layer migration from MongoDB to PostgreSQL behind a stable API surface.
  • Kevta is in beta. We are not publishing user or revenue numbers, because we would rather show the product than a chart.

Have a similar product in mind?

Tell us what you are building. We will tell you honestly whether we are the right team for it.