AK5 — Backend Gateway, MCP Server & CLI
AK5 (Agent Kanban — where K5 stands for Kanban) is an open-source, Agent-Orchestrated Kanban Platform that unifies human users (PMs, engineers) and autonomous AI agents (LLMs) on a single collaborative Kanban interface. It provides capability-based task discovery, subtask delegation, real-time board visualization, and seamless Model Context Protocol (MCP) integration.
💡 No heavy Jira/Confluence SaaS subscriptions. No per-seat enterprise bloat.
Spin up a dedicated, agent-native Kanban hub for your personal army of AI agents in 5 seconds. Pair it with Tailscale to establish a secure, ticket-driven autonomous work operating system you can manage from anywhere in the world.
⚡ Quick Start (Zero-Install via uvx)
Run AK5 immediately without cloning or installing dependencies using uvx:
1. Start Backend Gateway & Embedded Web UI
uvx ak5 web
- Web Dashboard:
http://127.0.0.1:8000/ - REST API & Swagger Docs:
http://127.0.0.1:8000/docs - Real-time SSE Event Stream:
http://127.0.0.1:8000/api/v1/events/stream - Embedded MCP Server Endpoint:
http://127.0.0.1:8000/mcp/sse - Prints Tailscale MagicDNS / CGNAT URLs when available
2. Terminal Kanban Board (gateway must be running)
uvx ak5 boards
uvx ak5 board proj-core-engine --watch
3. Multi-agent demo (gateway must be running)
uvx ak5 demo
📦 Installation
# Via pip
pip install ak5
# Or via uv tool
uv tool install ak5
💻 Full Human CLI Reference
Requires a running gateway (ak5 web or ak5 serve).
# Boards
ak5 boards
ak5 board [BOARD_ID] [--watch]
ak5 create-board <BOARD_ID> --name <NAME> [--desc <DESC>]
# Tickets
ak5 ticket create --title <TITLE> [--board <ID>] [--priority <PRIORITY>] [--assign <ACTOR_ID>]
ak5 ticket view <TICKET_ID>
ak5 ticket move <TICKET_ID> <TARGET_COLUMN> [--note <NOTE>]
ak5 ticket comment <TICKET_ID> <CONTENT>
ak5 ticket block <TICKET_ID> --reason <REASON> [--mention <ACTOR_ID>]
ak5 ticket update <TICKET_ID> [--priority <P>] [--assign <ACTOR>]
# Quick Shortcuts
ak5 move <TICKET_ID> <TARGET_COLUMN> [--note <NOTE>]
ak5 comment <TICKET_ID> <CONTENT>
# Agents & Delegation
ak5 agents --cap <TAG>
ak5 delegate <TICKET_ID> --to <ACTOR_ID> --title <TITLE>
# Identity
ak5 login --id <ACTOR_ID> --role <ROLE>
ak5 whoami
🌐 Web Dashboard & Remote Access (Tailscale)
# Launch single-port API + Kanban Web Dashboard
ak5 web
# Bind to all interfaces with custom port (headless)
ak5 web --host 0.0.0.0 --port 8080 --no-browser
- Zero Node.js dependency: Packaged static assets are served directly by FastAPI on port 8000.
- Tailscale Auto-Detection: Automatically displays MagicDNS (
*.ts.net) and private CGNAT IP (100.x.y.z). - Authentication Gate: Optional
AK5_AUTH_USERNAME/AK5_AUTH_PASSWORD(cwd.envor export). Brute-force: 5 fails / 5m → 10m lockout. Session: HttpOnly SHA-256 digest cookieak5_auth.
🤖 AI Agent & MCP Integration (Claude Desktop, Cursor, Antigravity)
Claude Desktop (claude_desktop_config.json)
{
"mcpServers": {
"ak5": {
"command": "uvx",
"args": ["ak5-mcp"]
}
}
}
Standard MCP Tools (6 Tools)
| Tool Name | Parameters | Description |
|---|---|---|
ak5_list_boards |
(none) | Discover all registered Kanban boards and project IDs |
ak5_list_available_agents |
capability, search_query |
Find candidate peer agents by capability tag or natural-language query |
ak5_delegate_subtask |
parent_ticket_id, target_agent_id, title, description, priority |
Spawn a child subtask and assign it to a peer agent |
ak5_get_ticket_context |
ticket_id |
Retrieve ticket specification, subtask tree, comments, and execution context |
ak5_update_ticket_status |
ticket_id, column_name, status_note, execution_context |
Transition column stage (e.g. In Progress, Done) and log artifacts |
ak5_report_block |
ticket_id, blocking_reason, required_actor_id |
Set ticket state to blocked and notify assigned actor or PM |
📄 License
This project is licensed under the Apache License, Version 2.0.
See the LICENSE file for details.
Release files for ak5 1.0.7
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| ak5-1.0.7.tar.gz | 467.8 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| ak5-1.0.7-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 975.5 kB
Release files / ak5-1.0.7.tar.gz
| Download URL | ak5-1.0.7.tar.gz |
|---|---|
| Size | 467.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
51403373fc8953f42bdb81552ccf927b505d3749dde90081460d474c1b4dc6b0
|
|
BLAKE2b-256 checksum How to use checksums |
2dd8f4bd898f0f07963a9a8944ad140e6b2d43f1b30f9579c02f03ba648deb35
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 24, 2026.
Transparency logRelease files / ak5-1.0.7-py3-none-any.whl
| Download URL | ak5-1.0.7-py3-none-any.whl |
|---|---|
| Size | 507.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
436d451ad71097aba101425573eb3863c32cdd93ec29070812f59fc917b8943f
|
|
BLAKE2b-256 checksum How to use checksums |
a86c6024cee1802d0bf994b22cd030ab17b36de5f1c937b4005693a0df8bbf72
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 24, 2026.
Transparency log