Plans, boards, and tasks for your team.

Track what everyone’s working on, move tasks across a board, leave comments, search across everything. Nothing you haven’t seen before — just clean, fast, and yours.

What’s inside

Everything you’d expect — and nothing you wouldn’t.

Boards & tasks

Kanban boards with drag-and-drop, nested tasks, labels, and optimistic concurrency. Everything scopes to your organization by default.

Keyword search

Meilisearch out of the box. Per-tenant scoped keys so the browser can query directly — without leaking a master key or crossing a tenant line.

Audit everything

Every mutation lands in an append-only audit log with actor, resource, diff, and request ID. Invite-only registration; admin bootstrapped from env on first boot.

Comments & attachments

Talk through the work next to it. @-mention teammates, drop files up to 100 MiB, edit history kept. Markdown-friendly rich text.

How it’s built

Four rules that keep the project boring on purpose.

Simplest stack that works: pnpm + Next.js + Postgres + Drizzle. No microservices, no Redis, no worker process — until the load actually requires one.
Zod is the source of truth: One schema per resource drives validation, OpenAPI, and the generated TypeScript client. Invalid input cannot cross the API boundary.
Tenant scoping in the repo layer: Every row carries organization_id. Repositories filter by the caller’s org context; the API layer cannot forget to pass it.
LLM-friendly REST: Stable prefixed IDs, problem+json errors, idempotency keys, and a condensed OpenAPI view — the Claude Code skill talks to it with curl.