Skip to content
Fredrin
Download
Documentation

Showing all documentation pages

Get started

Fredrin documentation

Everything you need to get started with Fredrin and understand how a board of AI coding agents fits together: the setup sequence from install to first merge, then the core concepts behind the board.

What is an agentic development environment?

An agentic development environment (ADE) is a workspace built for coding with AI agents at its center, where planning, execution, project context, and review all live on one surface instead of being scattered across a chat window, a terminal, and a separate issue tracker. It treats the agent as the primary way work gets done and wraps the whole loop, from an idea to a reviewed pull request, in a single tool.

Fredrin doesn’t compete with your agent: it works alongside Claude Code, Codex, and the shell you already use, and gives every ticket its own branch, worktree, and agent session. Instead of babysitting one terminal, you run a whole board of work in parallel and stay in command of what ships.

It is one shape among a wider field: see how the AI coding tools compare to place it against the editors, builders, and agents around it.

FIG 0.1 · One surface

Getting started

From install to your first merged pull request, in four steps.

FIG 0.5 · Install to merge

  1. 01Request an invite

    Fredrin is a native desktop app for macOS, Windows and Linux, invite-only while we scale, so the first step is to request an invite. Once you are in, the app is free to use, no credit card required.

  2. 02Install the app and connect a repository

    Install Fredrin and point it at the repository you want agents working in. The board you see is your backlog: every card on it is a ticket, and every ticket is one unit of work.

  3. 03Run your first ticket

    Create a ticket describing the change you want, then run it. Fredrin spins up a dedicated branch, a real git worktree, and an AI agent for that ticket, and streams the work back to the board.

  4. 04Review the diff and merge

    When the work lands in Review, read the diff. Ship it (the pull request and its CI status flow back onto the card automatically) or send it back with a comment and the agent picks up where it left off.

Core concepts

A handful of ideas explain how everything on the board fits together.

FIG 0.6 · Anatomy of a ticket
Ticket
A ticket is one unit of work. It runs through a single AI agent in its own git branch and worktree, and it carries its plan, its diff, and its pull request in one place. New tickets land in the backlog; running one spawns its agent.
Worker
A Worker is the AI agent session that builds a ticket. One ticket gets one Worker, so you can run many at once, each in its own lane, and throughput comes from parallelism rather than a single faster agent.
Worktree
Every ticket runs in its own real git worktree: a separate, isolated checkout of your codebase. Agents work side by side without overwriting one another, and you can check out any ticket's branch in a single click.
Project memory
Project memory is plain files in your repo: a wiki of concepts plus an append-only log of decisions, all cross-linked. Each Worker pulls the slice relevant to its ticket before it writes a line, so context compounds instead of scattering across one-off chats.
Bring your own subscription
Fredrin licenses the app, not the compute. Agents run on the AI coding subscription you already pay for (Claude, Codex, Cursor, or Copilot), with no per-token bill and no markup on model usage.

Guides

Step-by-step setup for the parts of Fredrin that touch your own machines and tools.

Go deeper

Each of these pages covers one part of Fredrin (or the git mechanics underneath it) in more depth:

The full documentation is still landing. The product it documents is already running.

Request inviteExplore features