Open source · Self-hosted · AGPL-3.0
Capture work in seconds.
Prove where the time went.
RyTask is an open-source project tracker for small, interrupt-driven teams — with native time tracking, first-class Slack capture, and full AI-agent control via MCP. Deep enough for engineers, friendly enough for everyone.
- 100% free core — no feature gates
- 49 MCP tools, full UI parity
- 1 Docker stack, your server
Why RyTask
Every team tool gets the same things wrong
The useful parts are paywalled
Slack capture and time tracking — the features small teams actually need — sit behind paid tiers everywhere else. In RyTask they're free, forever.
Your data lives in someone's cloud
The good tools are closed, capped, or cloud-only. RyTask runs on your server, and a full workspace export to JSON is one click away.
Self-hosted usually means heavyweight
Most self-hostable trackers are jargon-heavy and admin-hungry. RyTask uses plain language and sane defaults, so non-technical teammates actually use it.
AI agents are second-class
Other tools give agents a handful of read endpoints. RyTask gives them everything a person can do — enforced by an automated parity gate in CI.
Native time tracking
Honest time tracking, built in — not bolted on
A one-click timer in every task row, manual entries for when you forget, and every entry classified as planned work or an interruption. The plan-vs-actual report rolls it up per project and per person — with a full ledger, a “My week” view, and CSV export.
At the end of the week you can prove where the time went, not guess.
How the time model worksThis week · Website refresh
Fast capture
From Slack message to tracked task in under five seconds
Type one line — RyTask parses the assignee, estimate, and label, and the task is created, assigned, and tracked without anyone leaving the conversation. Message actions turn any existing thread into a task too.
The same grammar works everywhere you can type:
@assignee ~estimate #label !priority ^friday
Slack setup guide#support · Slack
AI-native
An MCP server with 100% workspace control
Anything a person can do in the UI, an AI agent can do over MCP — create projects, triage issues, log time, run reports. Parity isn't a promise: it's enforced by an automated gate in CI, served over streamable HTTP and stdio.
49 tools 100% parity, CI-enforced HTTP + stdio
Connect an agent{
"mcpServers": {
"rytask": {
"url": "https://rytask.your-team.com/mcp",
"headers": {
"Authorization": "Bearer <agent token>"
}
}
}
}Everything else
And the rest of a serious tracker
The core work loop
Projects, custom statuses, priorities, labels, sub-tasks, comments with mentions, saved views, full-text search, and a notification inbox.
GitHub linking
Branches, PRs, and commits link to tasks; magic words like fixes RY-12 auto-close on merge. Free from day one.
Real multi-tenancy
Organizations and workspaces, isolated by construction and verified by automated cross-tenant tests against real Postgres.
Data portability
Full workspace export to JSON — every entity. Your data is yours, and leaving is always allowed.
Non-technical-friendly
Plain language, sane defaults, zero jargon. Depth is available for the engineers — never imposed on everyone else.
Open by license
AGPL-3.0 — the whole core, including time tracking, Slack capture, and MCP. No open-core bait, no feature gates, ever.
Self-host
All you need is Docker
Three commands to a running stack — web app, API + MCP server, background worker, Postgres, and Redis. You land in a seeded demo workspace with the plan-vs-actual meter already doing its job.
git clone https://github.com/ali-maher-m/RyTask.git
cd RyTask
# The API refuses to boot without a strong JWT secret — generate one into a local override:
JWT_SECRET=$(openssl rand -hex 32) && cat > docker-compose.override.yml <<EOF
services:
api:
environment:
JWT_SECRET: "$JWT_SECRET"
worker:
environment:
JWT_SECRET: "$JWT_SECRET"
web:
build:
args:
NEXT_PUBLIC_API_URL: http://localhost:3001
EOF
docker compose up -d --buildThen open http://localhost:3000. For production, follow the self-hosting guide.
FAQ
Questions, answered
Is RyTask free?
Yes. RyTask's entire core is free and open source under the AGPL-3.0 license, with no paid tiers and no feature gates. Time tracking, Slack capture, GitHub linking, and the MCP server are all included at no cost.
Can I self-host RyTask?
Yes. RyTask is built to be self-hosted. It runs as a single Docker stack — web app, API and MCP server, background worker, PostgreSQL, and Redis — on your own server, so your data never leaves infrastructure you control. Three commands take you from a clone to a running instance.
Does RyTask track time?
Yes. Native time tracking is built in, not bolted on. Every task row has a one-click timer, you can add manual entries, and each entry is classified as planned work or an interruption. The plan-vs-actual report rolls time up per project and per person, with a ‘My week’ view and CSV export.
How does RyTask work with AI agents?
RyTask ships an MCP (Model Context Protocol) server that exposes 49 tools with 100% parity to the web UI — anything a person can do, an agent can do, including creating projects, triaging issues, logging time, and running reports. Parity is enforced by an automated gate in CI, and the server is available over streamable HTTP and stdio.
Does RyTask integrate with Slack?
Yes. You can turn a Slack message into a tracked task in under five seconds. A single line such as /task Fix the login bug @sam ~2h #bug is parsed for assignee, estimate, label, priority, and due date, and message actions convert any existing thread into a task without leaving the conversation.
How is RyTask different from Jira, Linear, or Asana?
RyTask is free and open source, self-hosted, and includes the features those tools paywall or omit. Time tracking and Slack capture are part of the free core, AI agents get full MCP control rather than a few read-only endpoints, and the interface stays friendly enough for non-technical teammates while remaining deep enough for engineers.
What license is RyTask under?
RyTask is licensed under AGPL-3.0-only. The whole core — including time tracking, Slack capture, and the MCP server — is covered, with no open-core carve-outs and no feature gates.
Can I export my data?
Yes. A one-click workspace export produces a complete JSON file covering every entity, so you can back up or leave at any time. Workspaces are multi-tenant and isolated by construction.
Own your tracker. Keep your time.
Free and open source under AGPL-3.0 — built for small teams whose week never goes to plan.