Feature

AI Task Auto-Execution

Move a task to "In Progress" and AI writes the code, commits it, and opens a Pull Request in your GitHub repo — all automatically.

End-to-End Execution Flow

1

Trigger Execution

Drag a task from Backlog to "In Progress" on the Kanban board. This is the only action you need to take — everything else is automatic.

2

Task Queued & Dispatched

A TaskExecution record is created and dispatched to a Node.js worker. Tasks within the same project are queued one at a time to avoid conflicts.

3

Clone & Branch

The worker clones your linked GitHub repository and creates a new branch named task/{id}-{slug} based on your default branch.

4

AI Writes the Code

The AI agent reads your codebase, understands the project structure, and writes the code. It can read files, edit files, run bash commands, and more — just like a real developer.

5

Commit, Push & Pull Request

Once done, the worker commits all changes, pushes the branch to GitHub, and creates a Pull Request with a summary of what was done.

6

Review & Merge

Review the PR on GitHub. When satisfied, move the task to "Completed" on the Kanban board and the PR is automatically merged.

Real-Time Progress Tracking

While the AI is executing, you see live updates in the execution panel:

Progress log — each step the AI takes (reading files, writing code, running commands)
Files changed — a list of files created or modified during execution
Token usage & cost — track input/output tokens and estimated cost
PR link — direct link to the Pull Request once created

What You Need

  1. A Premium subscription (or AppSumo lifetime deal).
  2. A GitHub account connected via OAuth.
  3. A GitHub repo linked to your project.
  4. At least one AI provider API key added (Anthropic, Copilot, DeepSeek, or Kimi).