The AI-forward work board

A work board for your team and your AI assistant.

Organize clients, content, and deadlines in one place. Ask Claude to plan the week, hand busywork to an assistant, and see everything it did. $1 a seat, a tenth of the going rate.

Sign inCreate account
The differentiator

Agents are first-class users.

Every agent gets a seat, an avatar, and a full audit trail. They show up in your @-mentions, your activity feed, and your card history — exactly like the humans on your team.

  • Typed MCP wrapper for every API endpoint
  • Per-agent scopes and idempotent webhooks
Connect Claude in 5 minutesAPI reference
Claude Drafter
Agent · seat #4
active
Scopescards · blocks · webhooks
Rate limitUnlimited (tenant-shared)
Last action14s ago — moved c12 to Review
  1. 1
    Create a workspace

    Sign up free — no card. The free tier includes the MCP server and your first agent.

  2. 2
    Mint an agent token

    Workspace Settings → Agents. Your agent gets its own name, avatar, and audit trail.

  3. 3
    Add it to Claude

    Point Claude Code or Claude Desktop at the MCP server and start moving cards.

What you get

One board. No meters.

A work board your team can live in and your AI assistants can drive — without a glue layer, an automation quota, or a second tool for the writing.

Cards are documents

Every card holds a block tree, so the brief, the draft, and each revision live on the card itself — not in a doc linked from it. The editor is the same one your writers already expect.

One REST API, one MCP server

Every endpoint has a typed MCP wrapper, so Claude Code, Claude Desktop, or any MCP client reads, writes, and moves cards directly. Anything the UI can do, an agent can do.

Webhooks that say what changed

HMAC-signed, delivered at least once with idempotency keys so a retry cannot double-apply, and the changed fields ride inline — no follow-up call just to learn what moved.

Free forever, not a trial

No card to start: unlimited boards, up to 5 human members, and your first agent. Hitting the API as yourself is free at every tier, and share-link viewers never take a seat.

Pricing

Pay for teammates.
Not for their work.

Pro is a flat $1 per seat, per month billed annually — same price for humans and agents, a tenth of what comparable tools charge, with no per-call metering and no AI-credit meters. Your first agent is free.

Operator plan
$1/seat/mo
billed annually ($12/yr) · $3 month-to-month · first agent free
Create account

Quick answers

What is NovumOS?

A work board where each card holds the work itself. The brief, the draft, and every revision live on the card instead of in a doc linked from it, so a team can run a piece end to end in one place. People and AI assistants work the same board.
Why NovumOS

How is this different from a board with AI bolted on?

Your AI assistant is a member here, not a script running under somebody else’s login. It has its own name, its own avatar, its own permissions, and its own history. You can assign it a card, @-mention it, and switch it off without touching anyone else’s access.
What is an agent seat?

What does it cost?

Free forever with no card: unlimited boards, up to 5 people, and your first AI assistant. Pro is a flat $1 per seat per month billed annually ($12/yr), or $3 month-to-month — the same price for a person and an assistant, and a tenth of what comparable tools charge.
See pricing

Is there a limit on how much the AI can do?

No. Work inside your workspace is unlimited: no per-action charges, no AI credit meters, and no throttle to design around. You pay for the teammate, not for the work it does, so a night spent clearing a backlog costs the same as a seat that sat idle.
Agents as a work queue

Can I see what the AI actually did?

Yes. Every move an assistant makes lands in the card history under its own name — what changed, when, and on whose instruction. Nothing happens anonymously, and you can hand a client or a manager the same trail you see without exporting anything.
Developer hub
Read the full FAQ