Skip to main content
dotbot is a PowerShell-native framework that brings structure and accountability to AI-assisted software development. Rather than generating code in a black box, dotbot runs each task through a transparent, version-controlled pipeline where your team can see what was decided, why, and how to pick up where it left off. Every AI session, question, answer, and code change is tracked directly in your repository.

Quick Start

Install dotbot and run your first AI-assisted workflow in five steps.

Installation

Install on Windows, macOS, or Linux via Homebrew, Scoop, or a source checkout.

Core Concepts

Understand workflows, stacks, task types, and the two-phase execution model.

CLI Reference

Full reference for every dotbot CLI command and flag.

MCP Server

Connect Claude, Warp, and other AI tools to dotbot’s 31-tool MCP server.

Dashboard

Monitor workflows, inspect audit trails, and manage settings from the web UI.

What dotbot does

Most AI coding tools produce output without a record of how they got there. dotbot solves this by wrapping AI execution in a managed workflow engine where every task runs in isolation and every decision is captured.

Workflow pipelines

Define multi-step AI pipelines in workflow.json with tasks, dependencies, and form configuration.

Two-phase execution

Analysis resolves ambiguity and builds context. Implementation writes code. No guesswork.

Git worktree isolation

Every task runs on its own branch. Parallel tasks never conflict.

Multi-provider AI

Switch between Claude, Codex, and Antigravity. Set models per task to control cost.

Enterprise registries

Publish and share workflows across your organization via git-hosted registries.

Human-in-the-loop

Route AI questions to stakeholders via Teams, Email, or Jira when input is needed.

Get started in three steps

1

Install dotbot

Use a package manager, which installs a self-contained copy and puts dotbot on your PATH:
Or install from a source checkout:
bootstrap.ps1 drops a PATH shim into ~/.local/bin (Linux/macOS) or %LOCALAPPDATA%\Microsoft\WindowsApps (Windows). Restart your terminal, then verify with dotbot status.
2

Initialize your project

Navigate to your project directory and run:
This creates a .bot/ directory with the tracked workspace/ tree and a .gitignore. The framework runtime is resolved from your active install, not copied into the project.
3

Launch the dashboard

Start the runtime and built-in web dashboard:
The dashboard opens at http://localhost:8686. From here you can run workflows, monitor tasks, and review session history. To start a workflow from the command line instead, run dotbot run start-from-prompt.
dotbot requires PowerShell 7+, Git, and at least one AI CLI - Claude, Codex, or Antigravity. No npm, pip, or Docker required.