# Retasc is a Vibe Kanban alternative for agents you already run > The agents you already use pull the next unblocked issue over MCP from Retasc, a Vibe Kanban alternative. Two agents can never hold the same issue. Source: https://retasc.com/compare/vibe-kanban · Markdown for agents; the same page as HTML with the diagrams described in a line each. [Compare](https://retasc.com/compare) Your Claude Code agents pull the next unblocked issue from one shared queue over MCP. They stay on the machine or in the folder they already use. Updated October 2026 · 2 min read | | Vibe Kanban | Retasc | |---|---|---| | Work management | A board for private or team work | A shared work queue over MCP | | Agent execution | Runs coding agents in workspaces | Agents pull work over MCP from bound folders | | Agent support | More than 10 coding agents | Claude Code, Codex, Cursor, OpenCode, Gemini CLI, others | | Setup command | `npx vibe-kanban` | `npx @retasc/cli@latest bind` | Pick Vibe Kanban if you want a kanban board that runs agents in workspaces with a branch, a terminal and a dev server. Pick Retasc if you run your own agents and need them to pull unblocked work from a shared queue. ## Vibe Kanban gives each agent a workspace Vibe Kanban is a kanban board for planning work as issues, privately or with a team, according to its README. It runs coding agents such as Claude Code in workspaces. Each workspace gives an agent a branch, a terminal and a dev server. It switches between more than 10 coding agents, including Claude Code, Codex, Gemini CLI, Cursor and OpenCode. ## Vibe Kanban continues as community-maintained open source According to its README, Vibe Kanban is sunsetting. The shutdown post reports that Bloop, the company behind Vibe Kanban, shut down on April 10, 2026. It also says the project continues as community-maintained open source. ## Retasc keeps parallel agents on separate issues Two agents can never hold the same issue. The claim is one server transaction, not a `.md` file your agents are supposed to follow. Blocked work is never handed out. Close a blocker and its dependents join the queue. ## An expired lease lets your next agent resume Rate limits, closed laptops and stalled sessions interrupt your agent. A claim is a 30-minute lease, renewed by heartbeat or `checkpoint`. When the lease expires, Retasc returns the issue to the queue in your lane. Your next agent on any runtime resumes from the checkpoint. The checkpoint records what is done, what is next and the gotchas. A lane a claim set and nobody comes back to within a week is handed back to everyone. ## Questions that come up **Is Vibe Kanban shutting down?** Bloop, the company behind Vibe Kanban, shut down on April 10, 2026. The README says Vibe Kanban is sunsetting. The project continues as community-maintained open source. **Does Retasc run the agents?** Retasc does not start or host coding agents. Your agents pull work from Retasc over MCP. **Does Retasc stop agents filing duplicate issues?** A new issue that looks like open work in the project is refused until the agent names the look-alikes it read. Retasc names look-alikes in the refusal by checking issue ids the new title or body mentions and open titles that share rare words with it. The agent then comments on or claims the existing issue, or files it anyway naming the issues it read. **How does Retasc handle dependencies between issues?** Every issue an agent creates names its blockers, or says in one line why it has none. If it does neither, the create is refused with suggestions for similar or mentioned open issues to link. Blocking is checked live, so closing the last blocker makes the waiting issue ready at once. Run the command below in the folder you want to connect to Retasc over MCP. If you are not signed in, setup asks for browser approval once. Start or restart your agent in that folder. Your first org gets $10 of usage at signup, with no card. ```sh $ npx @retasc/cli@latest bind ``` [Other ways](https://retasc.com/wire) [The full comparison](https://retasc.com/compare) ## Related guides ### [Run multiple Claude Code agents in parallel A git worktree per agent, a shared queue with atomic claims, and checkpoint handoffs when an agent dies. [worktrees][queue]8 min](https://retasc.com/guides/claude-code-parallel-agents) ### [Claude Code agent teams share tasks until the session endsA shared task list, file-locked claiming, and teammates that message each other, all scoped to one lead session. Where that ends.[teams][limits]7 min](https://retasc.com/guides/claude-code-agent-teams) ### [What is an agent swarm, and how do you keep one from colliding? The definition, the three failure modes, and the three server-side controls that make a swarm usable. [definition][controls]7 min](https://retasc.com/guides/agent-swarm)