Open-source project management and issue tracking, with AI agents built in

Self-hosted, open-source project management and issue tracking, and an alternative to Linear, Plane, and Jira. The difference: AI agents work here like any teammate. Hand them issues, or run the whole thing over the REST API, webhooks, and MCP.

If It's a Plan looks useful to you, star the repo ⭐ — it helps other people find it.

Website · Discussions · Contributing · Security

Most trackers bolt AI on as a chatbot. Here an agent is a real member of the team. It has a role, permissions, and an assignee slot, and it works the same board your people do.

It's a self-hosted, open alternative to Linear, Plane, and Jira. You run it on your own server and database.

• Own your stack. Your server, your database. No per-seat fees, no lock-in.
• Agents as teammates. Give an agent a model, a prompt, skills, and tools. Assign it issues like you would a person.
• One board for both. People and agents share the same board, threads, and cards.
• Automate anything. Drive it all through the REST API, MCP, and signed webhooks.

Heads-up: this is under active development. Expect breaking changes before the first stable release.

Tracking

• Configure issues per project: custom fields, labels, states, and issue types
• Kanban, table, timeline, and calendar views, with configurable display fields and two-level grouping
• Save any view as a reusable template
• Configurable dashboards for project analytics: throughput, breakdown, pulse
• Custom quick actions that run on an issue
• Freeform notes boards: sticky notes on a canvas, with colors, checklists, and connections
• Share a view or an issue by public link, read-only and without sign-in
• Initiatives that group and track work across projects
• Auto-archive, a notification inbox, role-based access control, and more

AI agents

• Agents as project members with their own permissions and assigned issues
• Configure each agent's model, system prompt, tools, and reusable skills — built on the Mastra agent framework
• Mention an agent in a comment to trigger a run
• Scheduled agent runs
• Built-in chat with per-agent conversation history

Platform

• REST API with an OpenAPI reference and API keys
• MCP server, so an external assistant can read and change issues through the same API
• Outgoing webhooks: subscribe to events, signed payloads, and retries with a delivery log
• Email and password, passkey, and Google sign-in
• Notifications by email (SMTP or Resend) and Telegram, with per-member preferences
• Instance administration: storage limits, mail transport, and instance-wide settings

Requirements: Bun 1.3+, Docker.

bun install
cp .env.example .env
cp apps/web/.env.example apps/web/.env

docker compose -f docker-compose.dev.yml up -d   # Postgres + MinIO only
bun run db:migrate
bun run dev                                      # api + web together, via Turborepo

api on http://localhost:3000, web on http://localhost:3001. bun run dev runs the whole workspace in watch mode from the repo root; the dev compose brings up only the backing services and the apps run on the host.

If host port 5432 is taken, set another and update DATABASE_URL in .env:

POSTGRES_PORT=5433 docker compose -f docker-compose.dev.yml up -d

Tests run against a real test Postgres, not mocks. Prepare a dedicated *_test database once, then run the suite from the repo root:

cp .env.test.example .env.test   # DATABASE_URL must name a *_test database
bun run db:migrate:test          # migrate the test database
bun run test                     # run all suites via Turborepo

The dev compose (docker-compose.dev.yml) provides Postgres and MinIO for these tests; the attachments suite needs the MinIO bucket it creates.

Alternatively, run the same gate CI uses — the suite against a throwaway Postgres in a container built from the production image:

docker compose -f docker-compose.test.yml up --build \
  --abort-on-container-exit --exit-code-from api-test

Point a Docker Compose resource at docker-compose.coolify.yml. Coolify supplies the secrets and domains through its own generated variables; nothing else to configure.

Requirements: Docker and a domain behind a TLS-terminating reverse proxy.

git clone https://github.com/croffasia/itsaplan.git
cd itsaplan
cp .env.example .env
# Set the public origins: API_URL, APP_URL
# Generate each secret with `openssl rand -base64 32`:
#   POSTGRES_PASSWORD, BETTER_AUTH_SECRET, APP_ENCRYPTION_KEY,
#   WORKER_INTERNAL_TOKEN, S3_ACCESS_KEY_ID, S3_SECRET_ACCESS_KEY

docker compose up -d --build

One command brings up the whole stack: Postgres, MinIO, api, worker, bot, and web. Every service is built from this checkout; web bakes API_URL into its bundle because Next.js inlines NEXT_PUBLIC_* at build time. The API applies migrations on startup, and the first account registered becomes the instance admin.

After an update, rebuild: git pull then docker compose up -d --build.

CONTRIBUTING.md covers the commands, the layout, and the conventions.

apps/api        Elysia HTTP API, mounts better-auth at /api/auth/*
apps/web        Next.js app, server-side rendered
apps/worker     webhooks, notifications, agent runs
apps/bot        Telegram bot
packages/db     Drizzle client, schema, migrations
packages/auth   better-auth server instance
packages/crypto AES-256-GCM encryption for secrets at rest
packages/mailer SMTP and Resend transport

The web app never imports the packages directly, it talks to the API over HTTP.

Vibe Code Kit — a Claude Code plugin with 20+ expert skills that turn AI slop into senior-level code. It teaches your AI agent professional development and design practices, so you ship production-ready code on the first try. Works with Claude Code, Codex, Gemini CLI, and Cursor.

It's a Plan is built with it — the code you are reading is the proof.

Issues and pull requests are welcome — bug fixes, features, docs, all of it. Start with CONTRIBUTING.md. It covers the setup, the conventions, and how a change gets merged. Be kind; we follow a Code of Conduct.

Not ready to contribute code? Star the repo and share it — it is the simplest way to help the project grow.

Found a vulnerability? Report it privately through GitHub Security Advisories, not a public issue, so we can fix it first. Details in SECURITY.md.

Copyright © 2026 VIBE DEV SPACE LLC.

AGPL-3.0. Running a modified version as a network service means publishing your changes under the same license.