Installation & Setup Usage Guide Configuration Agents & Tools MCP Servers Session Sharing ACP Integration System Architecture Embedding Guide API Reference Overview Configuration Caching Overview Payments Integration Guide OpenClaw Plugin

Getting Started

Install otto and start coding with AI in under a minute.

Install

Recommended: One-Liner

curl -fsSL https://install.ottocode.io | sh

Detects your OS and architecture, downloads the prebuilt binary, and installs to ~/.local/bin.

Pin a specific version:

OTTO_VERSION=v0.1.175 curl -fsSL https://install.ottocode.io | sh

Alternative: npm or Bun

bun install -g @ottocode/install

The postinstall script downloads the correct binary for your platform.

Supported platforms: macOS (x64, ARM64), Linux (x64, ARM64)

From Source

Requires Bun v1.0+.

git clone https://github.com/nitishxyz/otto.git
cd otto
bun install
bun run compile    # builds to dist/otto

Setup

1. Configure a Provider

otto setup

Walks you through provider selection and authentication interactively.

Or set API keys via environment variables:

export ANTHROPIC_API_KEY="sk-ant-..."
export OPENAI_API_KEY="sk-..."
export GOOGLE_GENERATIVE_AI_API_KEY="..."
export OPENROUTER_API_KEY="sk-or-..."

2. Start Using otto

otto                           # start interactive TUI (default)
otto --web                     # start server + web UI (opens browser)
otto ask "explain this error"  # one-shot question
otto ask "write tests" --agent build
otto ask "follow up" --last    # continue last session

3. Verify Installation

otto --version                 # check version
otto doctor                    # check configuration
otto agents                    # list available agents
otto models                    # list available models

How It Works

When you run otto, it:

Starts a local HTTP server for the otto API
Launches the interactive TUI in your terminal
With otto --web or otto serve, opens the web UI in your browser

All AI interactions, session storage, and tool execution happen locally on your machine.

For one-shot usage (otto ask "question"), it starts the server in the background, sends the prompt, streams the response, and exits.

Server Mode

otto serve                     # start on a random port, open browser
otto serve --port 3000         # specific port
otto serve --network           # bind to 0.0.0.0 for LAN access
otto serve --no-open           # don't open browser

The server exposes:

API on the specified port (e.g., http://localhost:3000)
Web UI on port + 1 (e.g., http://localhost:3001)

Troubleshooting

`otto` not found after installation

echo $PATH | tr ':' '\n' | grep local

# If not present, add it:
echo 'export PATH="$HOME/.local/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc

Provider authentication issues

otto auth login                # reconfigure credentials
otto doctor                    # check what's configured

Binary not executable

chmod +x $(which otto)