read
Opens your files and uses what's in them as context for the job.
OpenGUI reads your files, makes the changes, and runs the commands, so the job actually gets done. Bring your own model. No Git or coding-agent CLI to install.
Free and open source. macOS, Windows, and Linux.
Demo job: Add a spring sale banner to the homepage with a countdown to March 31. OpenGUI uses its tools on read index.html, read styles.css, edited index.html, created countdown.js, edited styles.css, ran npm run build, then replies: Added a sale banner above the hero with a live countdown to March 31. What changed index.html banner markup with a link to the sale page countdown.js updates every second and hides the banner when the sale ends styles.css banner colors reuse your existing button styles The build passed. Open `dist/index.html` to preview it.
OpenGUI is not a window on top of another tool. It does not run any of these behind the scenes:
Its own Harness does the work, saves every session's history, and uses exactly four tools in the folder you choose.
readOpens your files and uses what's in them as context for the job.
writeCreates new files: drafts, scripts, pages, slides, or plans.
editChanges one exact part of a file and leaves the rest alone.
shellRuns non-interactive commands in the project folder, like builds and checks.
Pick the installer for macOS, Windows, or Linux and open the app. It's free and open source.
Add the base URL, API key, and model name for any OpenAI-compatible endpoint. That's the one thing OpenGUI needs from you.
Start a session, describe the job, and watch each tool call and result as OpenGUI works.
Connect the desktop app, a browser, or your phone to the same OpenGUI Host and the same Sessions are waiting, with transcripts, model choice, and queued prompts included. On its own, the desktop app runs a Local Host on your computer.
Read the technical overviewInstall the app, connect your model, and hand it the job you've been putting off.