Getting Started

Getting Started

1. Install

Download the latest release from GitHub Releases:

Architecture	File
Apple Silicon (M1/M2/M3/M4)	omp-<version>-arm64.dmg
Intel	omp-<version>.dmg

Open the .dmg and drag omp into Applications.

The build is unsigned. On first launch, macOS may block it: right-click → Open, or go to System Settings → Privacy & Security → Open Anyway.

2. Sign In to a Provider

1. Open omp from Applications
2. Press ⌘, to open Settings
3. Navigate to Providers
4. Sign in via OAuth (browser flow) or paste an API key

Supported providers include OpenAI, Anthropic, Google, Grok, and 50+ others. Third-party providers can be added via API endpoint.

3. Start Your First Session

1. Press ⌘N to create a new session
2. Pick a workspace directory (the project you want the agent to work in)
3. Type a prompt and hit Enter
4. The agent streams its response — text, thinking blocks, and tool calls all render live

4. Explore the Command Palette

Press ⌘K to open the command palette. Every omp / command is available as a searchable, grouped menu:

• Workspace — directories, context, files
• Providers — login, models, API keys
• Session — new, switch, branch, handoff
• Tools — force tool, active tools, MCP servers
• Modes — plan, goal, vibe, loop

5. Go Parallel

• Press ⌥T to spawn a new tab with its own sidecar and session
• Each tab runs independently — background tabs keep running while you work elsewhere
• Optionally bind a tab to a git worktree (omp/gui/<name>) so parallel tabs never step on each other's files

Next Steps

• Features — explore the full feature set
• Keyboard Shortcuts — learn the keybindings
• Architecture — understand how it works under the hood