It's a Plan Launches Open-Source AI Issue Tracker

Overview

The open-source software landscape gained a significant new entrant with the release of It's a Plan, a project management and issue tracking platform designed to compete directly with established tools like Linear, Jira, and Plane. Unlike its commercial counterparts, It's a Plan distinguishes itself by treating artificial intelligence agents as first-class team members rather than mere add-ons.

The platform is built on a stack of modern, trusted technologies including Bun, Elysia, Next.js, and PostgreSQL, all containerized via Docker. Its core value proposition centers on data sovereignty and flexibility: users can self-host the application on their own servers, utilizing their own databases and AI model keys. This approach eliminates per-seat licensing fees and ensures that sensitive project data never leaves the user's infrastructure.

Released under the AGPL-3.0 license, the project invites developers to deploy it via platforms like Railway or through custom installations. By bringing your own model keys, organizations can integrate existing AI providers without committing to a proprietary ecosystem, aligning technical execution with enterprise-grade security requirements.

Core Philosophy: AI Agents as Teammates

The defining feature of It's a Plan is its architecture for integrating AI agents into the daily workflow of a development team. The platform posits that agents should not be isolated bots but collaborative entities that share the same board as human contributors. This integration allows for a seamless division of labor where human oversight coexists with automated execution.

The Collaborative Board

On the It's a Plan board, issues are treated as shared workspaces for both people and machines. A human developer can assign an issue to an agent or mention it in a comment thread, prompting the agent to begin work immediately. The agent then operates autonomously, posting replies in the thread, opening linked issues, and moving cards across the board just as a human would.

This interaction is exemplified in recent demo scenarios where a user named Mira Vos delegated an issue to a Triage Bot. The bot reviewed the importer, identified a rate-limiting issue, and autonomously created a follow-up task (PLAN-131) while adding backoff logic to webhook retries. This demonstrates how agents can handle routine triage and technical debt detection without constant human intervention.

Agent Configuration

Setting up an agent is designed to mirror the onboarding of a new employee. Users define a role, assign necessary skills, and configure tools that the agent can access. The platform supports two primary modes of operation:

Hosted Agents: These run within the application infrastructure. Users select a model provider, input their API key, and define system prompts. The agent is immediately available for assignment.
External Agents: For tasks requiring local codebases or specific CLI tools, users can deploy an agent runner on their own machine or VPS. This allows agents to interact with local repositories, logins, and development environments that the hosted version cannot access.

Technical Stack and Self-Hosting

It's a Plan is built for organizations that demand control over their technology stack. The application is fully containerized and relies on a robust set of open-source components:

Backend: Elysia.js running on Bun for high-performance request handling.
Frontend: Next.js for a responsive and dynamic user interface.
Database: PostgreSQL managed through Drizzle ORM, ensuring type safety and reliable migrations.
Storage: MinIO for object storage needs.
Containerization: Docker support for simplified deployment and scaling.

"Keep control of your stack. Run it on your own server, database, and model. Your data stays with you. No per-seat fees." — It's a Plan documentation

This self-hosted model addresses growing concerns regarding data privacy and vendor lock-in. By allowing users to bring their own AI model keys, the platform avoids mandating a specific provider, thereby supporting a wide range of OpenAI-compatible endpoints and local model deployments. The REST API and MCP (Model Context Protocol) support further enable automation, allowing teams to drive the tracker through custom scripts and external tools.

Features and User Interface

The interface of It's a Plan is designed to be familiar to users of Linear and Jira, yet offers greater customization and agent-specific functionalities. The board view supports multiple layouts, including Kanban, Table, Timeline, and Calendar views. Users can filter issues on any field, shape column layouts, and save these configurations as reusable views.

Key features include:

Cycles and Sprints: Built-in support for agile workflows, allowing teams to organize work into time-boxed iterations.
Custom Fields: Teams can define issue-type-specific fields to track metadata relevant to their domain.
Import Capabilities: Backlogs can be imported from CSV files, facilitating migration from other tools.
Permissions: A granular role and permission matrix ensures that access control can be tailored to project needs.
Webhooks and API: Signed webhooks with retry logic and scoped API tokens provide secure integration points for external systems.

The platform also introduces Collections for saved views and Notes Canvas with autosave functionality, enhancing the collaborative experience. The ability to bulk-edit issues and schedule agent runs via cron jobs further underscores the tool's focus on efficiency and automation.

Market Position and Availability

It's a Plan enters a crowded market dominated by Jira and Linear, but it carves out a niche by targeting organizations that require self-hosting and AI-native workflows. While Linear is praised for its speed and design, and Jira for its enterprise depth, It's a Plan offers a hybrid approach that combines the agility of modern trackers with the autonomy of open-source software.

The project is currently available for exploration at itsaplan.dev. Early adopters can deploy it immediately and begin configuring agents to handle tasks ranging from triage and code review to documentation updates. As the platform evolves, with planned features like live board refreshes and advanced permission matrices, it aims to become a viable alternative for teams seeking to integrate AI agents into their project management processes without sacrificing data privacy.