How it works

Capture, organize, automate, ship.

Projoint runs the loop your team already runs — and puts AI inside it as a teammate that proposes, never a black box that surprises you.

01

Capture

Type a thought in plain English. Cmd+K turns "ship the v1 landing by Friday, high priority" into a structured task — owner, due date, tags, and project, all filled in.

02

Organize

Tasks roll up into projects, boards, and timelines. Chat, files, docs, and the time tracker all live next to the work — not in five other tabs.

03

Automate

Agents watch for triggers — a PR opens, a task stalls, a morning arrives — and propose the next move. You stay in control: propose → approve → execute.

04

Ship

Approvals, an immutable audit trail, and live reports keep everyone aligned. The work moves forward; the record writes itself.

✨ Human-in-the-loop

The Trust Ladder — how much an agent can do is your call.

Most "autonomous AI" asks you to trust it blindly on day one. Projoint starts every agent at the bottom rung and lets you raise it — per agent, per scope — as it earns it. You can always climb back down.

L1

Suggest

The agent drafts — a task, a reply, a summary — and leaves it for you. Nothing changes until you say so. The safe default for a new agent.

L2

Propose

The agent prepares a concrete change and asks for approval in one click. You see exactly what will happen before it does.

L3

Act in scope

Inside a project or task type you have whitelisted, the agent executes directly — and writes every action to the audit log. Trust earned, scope bounded.

L4

Act freely

Full write access for a mature, trusted agent. Server-enforced, budget-capped, fully logged — and revocable in one toggle the moment you want it back.

Governance is the feature, not the fine print.

The same controls that govern your people govern your agents — because to Projoint, an agent is just another team member with a role.

Roles & permissions

Custom roles decide who — human or agent — can create, edit, approve, or administer. Permissions are checked at the database, not just the UI.

Every action audited

Status changes, approvals, agent runs, role grants — written to an append-only log that even an admin cannot rewrite.

Budgets & kill-switch

Each agent runs against a token budget. One admin toggle disables all AI instantly — no redeploy, no waiting.

Ship the next thing faster.

Free for ten users. No card required. AI optional.