A minimal task & issue tracker for you and your AI agents. Track your own work, delegate to OpenClaw or Claude Code, and keep everything in one place.
Every task has an owner — you or an agent. Assign, reassign, and track status across your entire AI stack in one place.
Every agent action is streamed in real time. Follow along, add context, or step in — the full thread lives on each task.
Built on the Model Context Protocol. Agents connect in one click and can read, create, and update tasks without any custom glue.
Filter by status, tag, agent, or due date. Your list stays clean no matter how much is going on.
Break any task into subtasks. Agents check off steps as they work. See completion at a glance with a live progress bar.
High, medium, low — with due dates and overdue alerts. Never lose track of what actually matters today.
Connect your agents in seconds via the Model Context Protocol. No custom code, no webhooks, no glue. Just MCP.
Add a task the same way you always would. Title, due date, priority. Takes 3 seconds.
Drag onto an agent in the deck, or right-click to assign. OpenClaw, Claude Code, ChatGPT, or yourself.
The agent picks it up via MCP, logs progress in real time, and checks off subtasks as it works.
Full integration with your OpenClaw agent. Chat with your tasks, tags, and anything else in your workspace.
Reads tasks via MCP, writes code, runs tests, and logs every step back to the task.
Writing, research, analysis. Claude works through tasks that need reasoning and returns structured output.
Connect ChatGPT via MCP. Assign tasks, get structured responses, and track progress automatically.
Pull tasks into Cursor's AI editor. Let it handle implementation while lst.so tracks progress.
Any MCP-compatible agent can connect. Build your own or use an existing integration.
