VisionCLI sees your screen and uses your computer. Hold a key and ask about the file, error or page in front of you, or have it do the task: with Claude Code or Gemini, by voice or by typing.
Everything happens on your computer, with the AI you already use.
Hold ⌥ / Alt and ask out loud. Add ⌘ / Win to include a screenshot. The answer appears in a bubble and is read aloud.
"What is this file for?" It looks at the window under your mouse, finds the real file in your project and answers with file:line references.
Clicks, types and opens apps one checked step at a time. You approve each action, passwords always ask, and Esc stops it instantly.
Answers draw boxes, arrows and numbered steps on your real screen, and can include Mermaid diagrams.
"Open the Stripe tab", "find a stay in Lisbon with a pool": quick commands and multi-step tasks in Chrome, Brave and Edge.
Parakeet and Whisper transcription and Kokoro voices run on your machine. Your audio isn't sent anywhere.
Free and open source (MIT). Version 0.10.5.
brew install --cask thethirdsourcers/tap/visioncli
VisionCLI isn't notarized by Apple yet, so macOS asks before the first launch: System Settings → Privacy & Security → Open Anyway, or run xattr -dr com.apple.quarantine /Applications/VisionCLI.app.
winget install TheThirdSourcers.VisionCLI
Installs for your user only, no admin rights.
Debian / Ubuntu (apt repository, updates with the system):
curl -fsSL https://visioncli.thethirdsourcers.com/public.key | sudo gpg --dearmor -o /usr/share/keyrings/visioncli.gpg echo "deb [signed-by=/usr/share/keyrings/visioncli.gpg] https://visioncli.thethirdsourcers.com stable main" | sudo tee /etc/apt/sources.list.d/visioncli.list sudo apt update && sudo apt install visioncli-linux
Fedora / RHEL / openSUSE (dnf or zypper repository):
sudo curl -fsSL -o /etc/yum.repos.d/visioncli.repo https://visioncli.thethirdsourcers.com/visioncli.repo sudo dnf install visioncli-linux
Then start VisionCLI from your app menu; its setup wizard checks the desktop tools it needs. Details in docs/LINUX.md.
Windows releases: free code signing provided by SignPath.io, certificate by SignPath Foundation.
Every release is built by GitHub Actions from this public repository; nothing is built or signed on a personal machine. Committers and reviewers: the maintainers of thethirdsourcers on GitHub. Approver for each signed release: Manjit Shakya (Third Sourcers). See also the privacy policy.
Every file is listed in SHA256SUMS.txt and carries a build provenance attestation from the GitHub workflow that built it:
gh attestation verify VisionCLI-0.10.5-macos.dmg -R thethirdsourcers/visioncli