Embed your project management directly in your repository.

Stop wrestling with external issue trackers. Keep tasks, context, and a clear paper trail right alongside your code so human developers and AI agents can collaborate seamlessly.

Get Started Read the Docs

PROJ-1a2b3c

PROJ-4d5e6f

PROJ-7g8h9i

{ "id": "PROJ-1a2b3c", "title": "Calibrate flux capacitor", "type": "epic", "status": "backlog", "priority": 2, "assignee": "Codex" }

PROJ-1a2b3c

medium

Calibrate flux capacitor

Codex

PROJ-4d5e6f

high

Stabilize warp core coolant loop

Ryan

PROJ-7g8h9i

high

Diagnose tachyon scanner drift

Claude

Tracking project context usually means coordinating an external system. But it's better to eliminate a problem than to solve it. With Kanbus, every task is a file in your repo. Your workflow inherits the exact same version control as your code—creating a clear, timestamped paper trail without a jarring context switch.

Already using Beads? Your existing data works.

Use Git as a Kanban bus

Synchronize multiple boards, CLI tools, and AI agents through your repository's commit history. The repo is the message bus.

Git Synchronization

Realtime Kanban Board

Watch the board update in realtime as your agent makes changes.

Learn about Realtime Collaboration →

Backlog2

TSK-1a2b3c

medium

Calibrate flux capacitor

TSK-1a2b3c

low

Align temporal coils for flux channel

In Progress2

TSK-7g8h9i

high

Diagnose tachyon scanner drift

TSK-4d5e6f

high

Stabilize warp core coolant loop

ryan

Done1

TSK-0j1k2l

low

Ship holodeck safety interlocks

Select an issue to view its details.

How it works

Watch how Kanbus integrates project management directly into your terminal, editor, and agent workflows.

Features

Focused capabilities that make Kanbus practical for daily work.

Agent-Ready CLI→

A fast, scriptable CLI tool that lets your coding agents read requirements, update statuses, and track issues natively.

Kanban Board→

Visualize your workflow in a modern, reactive interface. Drag and drop issues, filter by status, and see the big picture.

Git-Native Storage→

Your issues are plain JSON files in your repo. No database, no proprietary index--just transparent files you can read, diff, and version like any other code.

Realtime Collaboration→

Instant updates across agents and humans. Lightweight pub/sub messaging keeps everyone in sync without waiting for a pull.

Jira Sync→

Pull Jira issues into your repository so agents always have the full context.

Local Tasks→

Keep private tasks off git and promote them when they are ready to share.

Virtual Projects→

Aggregate issues from multiple repositories into a single view.

Beads Mode→

Seamless compatibility with existing Beads projects. Zero config required.

VS Code Plugin→

Manage your project without leaving your code. Full Kanban board integration.

Integrated Wiki→

Generate dynamic documentation from your issues. Tables, queries, and live data—all from Markdown.

Policy as Code→

Combine hard guardrails with kairotic, in-the-moment guidance so agents follow procedure without losing flow.

Lifecycle Hooks→

Plug arbitrary project logic into Kanbus lifecycle events with before/after hooks and built-in policy guidance integration.

Agile Metrics→

Track filter-aware issue counts and status/type breakdowns in a focused metrics view built into the Kanbus console.

Issue Router→

Dispatch labeled issue packages through a coding agent (Codex or OpenCode) with deterministic capacity controls, Git-backed state, and a pull-request review gate.

Dual Implementation + Web Console

One behavior specification driving two complete CLIs, plus a web UI server.

Rust CLI

The preferred way to run Kanbus. High-performance binary for CI/CD pipelines and large repositories. Install via Cargo or download a pre-built binary.

Python CLI

Alternative implementation available via PyPi. Perfect for scripting, local workflows, and integrating with Python-based AI tools.

Web Console

Single-binary web UI server with embedded React frontend. Download and run—no configuration, no separate assets, no npm required.

Why Kanbus?

Built for the age of AI-assisted development.

Successor to Beads

Inspired by the elegant Beads cognitive framework, but re-engineered for Git. Kanbus drops the SQLite database, uses separate files to eliminate merge conflicts, and adopts standard Jira terms to better leverage AI pre-training.

Vs. Jira

Your data is local. No web latency, no login screens, and native CLI access for your AI agents to read and write tasks. And there are no per-seat costs—it's just your repo.

Vs. Markdown

You get structured data (priority, status) where you need it, and free-form Markdown descriptions where you want it.

Delegated responsibility

You can run Kanbus yourself. If you want help, Anthus can operate a shared setup, configure it with you, or adapt it to your team.

You choose how much we help

Fork it

Kanbus is MIT licensed. Run it in your own repository. Anthus does nothing and charges nothing.

No cost

Self-setup, managed

Set up Kanbus in your environment; we operate the shared services and keep the deployment current.

$20 a month

Assisted setup, managed

We configure Kanbus with you, then operate it and keep the deployment current.

$20 a month

and $100 once

Professional services

We adapt Kanbus workflows, integrations, or deployment to fit the way your team works.

Quoted

Straight answers

Will Kanbus always be free to run myself?

Yes. Kanbus is MIT licensed and has no per-seat license fee.

What does managed mean?

We operate the agreed shared services, apply updates, and help keep the deployment healthy.

Can I move between setup options?

Yes. You can begin with self-setup and add an assisted session later, or take over operation yourself.