Agent orchestration and automation

Run and automate work with AI agents.

Bring your agents and tools together. Start tasks, build repeatable workflows, and follow the results across your projects and machines.

  1. Idea
  2. Exploring
  3. Planning
  4. Ready to build
  5. Building
  6. Shipped
Works withClaude CodeCodexCursorAny MCP clientGitHubGitLab

Make useful workflows repeatable.

Create reusable workflows. Configure and save the steps you use for recurring work.

Keep the brief and parameters ready for the next task.

  • IDEAS
  • PLANS
  • DOCUMENTS
  • RUNS
  • DECISIONS
  • RELEASES
Product walkthrough

From an idea to
work in progress.

Capture what you want to do, turn it into a plan, and choose the agents that will work on it. Follow each run and review the results alongside the original request.

Example task

Add CSV export to the reports page, have another agent review it, and run the checks.

Implementation → Review → Checks

Explore the release workflow

Capture

Capture the idea and why it matters before work starts.

/capture
The capture modal

Plan

Start work from your plan. Choose an item or group of tasks, the agents, and the machine. Requirements and test notes stay in the brief.

/projects/kairokud/plan
The plan with the stage strip

Run

See what needs your attention. Follow active runs, inspect results, and see where work has stopped.

/projects/kairokud/floor
The Floor with a run in flight

Ship

Keep release decisions with the work. A person decides what is done and what ships.

/projects/kairokud/releases
The Releases index with a shipped stamp
  • The workspace board
    /projects/kairokud/board

    Keep the reasoning with the result.

    Plans, research, decisions and run history stay with the project, so the next task starts with the work already done.

  • A daemon session claiming a run
    /daemon

    Use your agents on your machines.

    Choose the agent and execution environment for the job. Manage work across connected computers and servers from the web app.

Connected agents

Connect your agents to the work.

Give connected agents access to project plans and documents, and let them report progress back to Kairoku.

MCP tokens
/settings/tokens
Human review

Review before it ships.

Agents can implement changes and report their checks. You decide what to merge and when a release is ready.

A plan item's gate
/projects/kairokud/plan

Start with one task.

Kairoku is in early access and free while it is an MVP.

Connect a machine, choose a project, and run your first agent task in Kairoku.

Pricing

Free

Early access for individuals and small teams running work with AI agents.

$0while Kairoku is an MVP
  • Unlimited projects
  • Reusable workflows
  • Run the daemon on your own machine
  • Plans, runs and decisions together
Start free