AgentsInFlow
AgentsInFlow

Getting Started

Install and run AgentsInFlow, then capture and refine tickets quickly.

Prerequisites

AgentsInFlow supports three AI coding CLIs. Install at least one, then authenticate to start using it.

Claude Code

Anthropic
Install
npm install -g @anthropic-ai/claude-code
Login
claude login
View documentation →

Codex

OpenAI
Install
npm install -g @openai/codex
Login
codex login
View documentation →

Cursor

Cursor
Install
cursor-agent
Login
cursor-agent --login
View documentation →

AgentsInFlow auto-detects installed CLIs. You only need to log in once. Credentials persist across sessions.


Installation

1

Download the latest release from GitHub Releases

2

Open the DMG (macOS) or run the installer (Windows)

3

Drag to Applications (macOS) or follow the installer prompts (Windows)

4

Launch the application

Available for macOS (Apple Silicon) and Windows. Linux support is planned.


Windows Support

AgentsInFlow ships with a dedicated Windows installer and auto-updater. All features available on macOS are also available on Windows, including:

  • Full Kanban board, Git view, and Chat view
  • All three engine CLIs (Claude Code, Codex, Cursor)
  • Auto-update with background checks
  • Git worktree branching and terminal integration

First-Run Onboarding

When you launch AgentsInFlow for the first time, a guided setup flow walks you through the initial configuration:

1

Detect installed engine CLIs (Claude Code, Codex, Cursor)

2

Guide you to install missing CLIs with in-app install actions

3

Create your first project and start working


Updates

Auto-update checks are enabled in packaged builds.

  • First background check runs about 15 seconds after app launch.
  • Ongoing background checks run every 4 hours.
  • Manual check is available from AgentsInFlow → Check for Updates... on macOS.

Quick Start

01 Create a Project

Click "New Project" and select a folder. This folder should be a git repository. Choose your preferred default engine.

02 Configure Engine (Optional)

Open project settings to configure engine-specific options. These become the default configuration injected when you run tasks.

03 Create a Ticket

Click the "+" button on the Kanban board. Give it a title, description, and optionally set type and priority.

04 Run an Execution

Click a ticket to open the inspector. Click "Run", select your engine, model, and branching options, then click "Start". The CLI launches in an integrated terminal.


Quick Capture to Proper Ticket

Recommended flow:

1

Capture quickly with Command+G on macOS.

2

Run one exploration execution to collect context.

3

Ask assistant with app access to update the ticket description into a concrete plan.

4

Launch the next run using that refined ticket.