For teams leaving Vibe Kanban

Your board's home shut down. Your work doesn't have to.

Vibe Kanban proved developers wanted a place to orchestrate coding agents. Planwright is where that work goes next — objective-native, MCP-native, and built with the signed audit trail your old board never had. Bring your board over in a single command.

Migrate my board →What happened to Vibe Kanban?

Apache-2.0 verifier, open sourceWorks with Claude Code, Cursor, Codex, Amp, FactoryGitHub OAuth · no lock-in

The honest version

Vibe Kanban didn't fail on product. It ran out of business model.

On April 10, 2026, Bloop shut Vibe Kanban down — announced live on stage — with roughly 30,000 monthly active users and 26,000 GitHub stars. That is not a product that failed. The project lives on as a community-maintained, Apache-2.0, local-only tool, but the company behind it is gone.

The reason was economics, not engineering. As the founder put it, the only two ways to make money in AI coding tools today are selling to enterprise and reselling tokens — and Vibe Kanban was doing neither.

“We couldn't find a business model we could get excited about. Everyone making money is doing two things: selling to enterprise, and reselling tokens. We were doing neither.”— Louis Knight-Webb, on shutting down Vibe Kanban

We're not here to dunk on a good team. We're here because a few thousand engineers just lost the tool they used to run their agents — and the thing that would have saved it is the exact thing Planwright is built around.

Why Planwright is the landing spot

The same agent workflow — on a foundation that lasts.

Planwright isn't a Vibe Kanban clone with a logo swap. It's the post-kanban model: you manage outcomes at the objective layer, and the agent does the breakdown and the work.

Objective-native

Not tasks. Not stories. Not points.

You write outcomes with acceptance criteria. Agents decompose and execute. No story points, no velocity, no capacity math — breakdown is the agent's job, not a board feature you maintain.

MCP-native

Every agent, day one.

A native MCP server, not a bolted-on integration. Claude Code, Cursor, Codex, Amp, and Factory claim objectives, post plans, record diffs, and request acceptance through one connection.

Audit-native

The evidence trail agents need.

Every objective and lane change is written to an append-only, hash-chained, ECDSA P-256-signed ledger — SOC 2 evidence for AI-generated code as a core primitive. The enterprise wedge Vibe Kanban never had.

The migration

Bring your board over — faithfully, not sloppily.

Install the onboarding plugin, point it at your Vibe Kanban export (or your Jira, Linear, Asana, or Monday board), and it maps your work into Planwright objectives while respecting the model. Here's exactly what carries and what doesn't.

What carries over

→Cards and tasks become atomic objectives — titles and descriptions intact
→Checklists and subtasks folded into acceptance criteria
→Priority mapped to Planwright priority
→Ready-for-dev items land in the scheduled lane, agent-pickup ready
→Everything scoped under a source label with a filtered board link

What we intentionally drop

×Story points, estimates, velocity, burndown — gone by design
×Sprint and iteration fields — scheduling is per-objective
×Bugs are separated out, never mixed into objectives
×“In progress” as a fake lane — real lanes come from agent claims
×Execution history from the old tool — your new trail starts clean and signed

# 1. Connect Planwright to your coding agent
claude mcp add planwright \
  --header "Authorization: Bearer $PLANWRIGHT_TOKEN" \
  https://mcp.planwright.tools/mcp

# 2. In Claude Desktop or Claude Code, just ask:
› "Migrate my Vibe Kanban board into Planwright"
› "Import EPIC-42 from Jira"   # or Linear / Asana / Monday

# The onboarding plugin previews the mapping, you approve,
# and it creates each objective through a signed MCP call.

Create your workspace Get the onboarding plugin

The part that pays the bills

An audit trail your CTO can hand to an auditor.

The 2026 AICPA Trust Services Criteria expect immutable, verifiable evidence for AI-generated code changes. A kanban board can't produce that. Planwright writes it on every agent action — hash-chained, signed in AWS KMS, and independently verifiable offline.

objective.created→agent_run.started→plan_appended→diff_recorded→acceptance.signed→✓ chain intact

Read the compliance brief →

Straight answers

What Vibe Kanban users ask first.

Do I lose my board data?+

No. Export your Vibe Kanban board (or read it from its local database), and the onboarding plugin maps it into Planwright objectives. It shows you the full mapping before it writes anything — you approve, then it creates each objective through the MCP.

Which coding agents work with Planwright?+

Claude Code, Cursor, Codex, Amp, and Factory — any MCP-compatible agent connects through one HTTP endpoint with a project token. Planning happens in Claude Desktop or claude.ai; execution happens in your IDE.

Why doesn't Planwright have story points or velocity?+

Because your agents don't need help breaking down work — they need a place to check in against outcomes. Points, velocity, and capacity math were built for humans decomposing tickets. Planwright is objective-native on purpose, and the migration drops those fields rather than pretending they still mean something.

What happens to my “Done” cards and my bugs?+

Completed work is left behind by default — your new signed trail starts clean rather than importing history that was never audited. Bugs are pulled into Planwright's separate bug surface, never mixed into the objective board. Both behaviors are deliberate.

Is there lock-in?+

Identity is GitHub OAuth, one org maps to one workspace, and the audit verifier is open source. Your evidence bundle is a portable, signed artifact you can verify without us. Leaving is always possible; the point is you won't want to.

Post-kanban is here

The control plane for autonomous software labor.

Get started free Migrate from Vibe Kanban