Installation

Requirements

• Python 3.11+ — check with python --version
• A Chromium-based browser — Chrome, Brave, Edge, Vivaldi, Opera, Arc, or Chromium
• macOS or Linux — desktop automation (computer) currently supports macOS only

Install

pip install omniscout

Verify the install:

omniscout --version
omniscout --help

Set up a browser

OmniScout drives a real browser two ways:

Check which browsers are installed and verify Chromium availability:

omniscout settings browsers
omniscout install

To drive your existing browser instead, point OmniScout at it and load the extension:

omniscout setup --browser dia

This downloads the extension and opens the browser's extensions page. Choose Load unpacked, select the printed folder, and confirm the popup says connected. See How it works for how the options differ.

Get ready

OmniScout stays ready so every action runs fast:

omniscout daemon start
omniscout daemon status

You only need to do this once — later commands get ready on their own. See Sessions for lifecycle and logs.

Get ready for search

Run once after install so the first search is fast:

omniscout warmup

Install agent skills (optional)

If an AI agent will drive OmniScout, install the skill files so Claude Code, Cursor, Codex, and Gemini CLI auto-discover it:

omniscout install --skill

See Using with AI agents for the full agent setup.

Update and uninstall

pip install --upgrade omniscout   # update
pip uninstall omniscout           # remove the package

To wipe all local state (profiles, caches, indexes, memory):

rm -rf ~/Library/Application\ Support/omniscout/   # macOS
rm -rf ~/.local/share/omniscout/                   # Linux

Or override the location with OMNISCOUT_DATA_DIR. See Configuration.

Next steps

• Quickstart — your first search, extract, and browser run in 5 minutes
• Troubleshooting — fixes for install and browser issues