Everything you need to run agents like a pipeline
Takuto turns a coding agent into a supervised, repeatable pipeline — with the controls, isolation, and visibility you would expect from your own infrastructure.
Run it your way
Choose how much autonomy each ticket gets — and switch whenever you like.
Autonomous mode
Connect Jira or GitHub Issues and Takuto polls automatically: it picks up "To Do" tickets, runs the full pipeline (worktree → install → implement → lint/tests → PR), and moves to the next one.
Manual mode, your pace
Add any ticket or task to the dashboard yourself, refine its description with AI assistance before the agent sees it, then trigger each phase when you are ready. No polling, no surprises.
Mixed mode
Auto-pick routine tasks while manually curating the tricky ones. You decide which tickets get the autopilot treatment and which you steer yourself.
Connect your work
Drive workflows from the tracker you already use — or from nothing at all.
Jira
Poll a Jira project for To Do tickets and patch descriptions as work progresses (requires Atlassian CLI auth).
GitHub Issues
Poll issues straight from your repository — the repo is detected from the cloned project’s git remote.
No ticketing
Paste any description into the dashboard and Takuto treats it as a workflow. No Jira account required.
Watch and steer
Full visibility into every run, and a way into any of them.
Parallel workflows
Configure how many workflows run at once; each gets its own git worktree and isolated environment.
Live dashboard
Stream terminal output per workflow, watch progress, and pause, resume, retry, or inspect any run in real time.
Browser editor + web terminal
Open a VS Code editor and web terminal, pre-configured with your project tools, pointed at the exact worktree the agent is working on.
Define the pipeline
Express your process as code, then trigger it from the dashboard.
Multi-step pipelines
Chain phases — implement, address PR comments, merge the base branch — with depends_on relationships between steps. Edit your workflows in the dashboard (Configuration → Workflows); the workflows/ directory seeds the defaults for a new workspace.
AI description refinement
Sharpen a ticket’s description with AI assistance before the agent ever runs, so the pipeline starts from a clear spec.
Multi-user, self-hosted
Deploy on a server so a whole team shares one dashboard, with each project fully isolated by its own volumes.
Want the details?
The documentation covers configuration, workflow definitions, and how a run executes end to end.