For teams building with agents
Install the workflow.
Watch your agent team run.
Onboarding, design, execution, review, ship. Every agent picks up the same one and writes to the same board, so the next one starts from what the team already decided. Rewrite any part of it.
Onboarding · Design · Execution · Review · Ship
30 days free · no card · one line to connect the agent you already use
Every agent starts from what the team already settled.
Watch one agent read the project, brief you, do the work, save what it discovers, and bring the decision it cannot make back to the board — where the rest of the team sees it, before it is code.
The board is where humans and agents meet.
Agents move work forward, record what changed, and raise decisions they cannot make. Everyone sees the whole project without reading each other’s chat logs or reconstructing what happened.
The method comes with the plugin.
Designing before it builds. Breaking that design into steps. Working the checklist, asking for a review, and landing the branch. It arrives as an agent plugin with skills, an MCP server and this board, and every piece of it is yours to rewrite.
How it works
Bring any coding agent onto the team.
Tuckit is an MCP server, not another chat window. Connect the agent you already use and it can read the project, do the work, and update the shared state without a separate handoff.
- 1Connect the agent you already runOne line in Claude Code, Cursor, Codex or Antigravity. You approve it in the browser; there is no key to paste anywhere.
- 2It reads the project before it writes codeWhat was decided, what that ruled out, and what is next — pulled at the start of the session instead of re-explained by you.
- 3What it learns lands where the team can see itProgress, discovered work and the choices it could not make go back onto the board, so the next person and the next agent both start there.
Run this, then click Allow in the browser tab that opens.
claude mcp add --transport http tuckit https://app.tuckit.dev/mcpA decision only one person can see is not a decision.
Your tracker holds what somebody wrote down afterwards. A markdown file lives in one checkout, and an agent’s memory lives in one agent. All three work until a second person starts building from something they never saw.
| Capability | An issue tracker | A markdown file | An agent memory server | tuckit |
|---|---|---|---|---|
| A teammate can open and read it | Yes | Partly | No | Yes |
| Survives a new session | Yes | Yes | Yes | Yes |
| Your agent writes to it while it works | No | No | Yes | Yes |
| Holds why it was decided, and what was ruled out | Partly | Partly | No | Yes |
| Every teammate’s agent reads the same state | Partly | No | No | Yes |
| Sprints, estimates and reports for people outside the team | Yes | No | No | No |
Pricing
$99 a month, per workspace.
One plan. Unlimited people, unlimited agents, unlimited history — a workspace costs the same with two people in it or twenty. Thirty days free to start, no card.
No card, and no countdown tricks. The thirty days belong to you rather than to the workspace, so a second workspace joins the same clock instead of starting a new one. If the price ever changes, everyone using tuckit hears it from us before it takes effect.
Use the cloud. Keep an exit.
Everything you and your agents write here is yours. Take the whole board out whenever you like — no ceremony, no support ticket, and nothing that only we can read.
Your team decides. Every agent starts there.
So nobody’s work gets thrown away because the decision was somewhere else.
Start free