Loading...
Product
Core Features
Top Extensions
Company
Read My Screen is a Raycast extension that turns screenshots, clipboard images, image files, and the active browser tab into AI-powered explanations — using OpenAI (GPT-4o and more), Anthropic Claude, or Google Gemini. You add your own API keys; requests go directly to the provider you choose (no custom backend).
Extension commands in Raycast.
Configure capture, model, and instructions before running analysis.
Analysis in progress with your chosen provider.
Structured reply, copy actions, and follow-up in the same thread.
Reopen past runs from Session History once you have saved conversations.
npm install
npm run dev
Build and lint:
npm run build
npm run lint
Versioning (Changesets): The semver lives in package.json (version). After a change worth releasing, run npm run changeset to record it, then npm run version-packages to bump the version and update CHANGELOG.md. Publish to the Raycast Store with npm run publish when ready.
Icon: The menu bar icon is generated from assets/extension-icon.svg. After editing the SVG, run npm run generate-icon (writes extension-icon.png for Raycast).
Open Read My Screen in Raycast → Extensions → Preferences to set API keys, default model (must match the provider whose key you use), and default instructions for new runs.
For reporting security issues in this codebase, see SECURITY.md.
MIT — see LICENSE.
Main command: capture your screen (region, window, full screen, or clipboard image) or read the current browser tab as text. Vision and text models; follow-up chat; session history.
One shortcut: analyze the image on your clipboard with your default model — shows the response immediately.
One shortcut: summarize or question the active browser tab (Chrome, Safari, Arc, Dia, Edge, and more) using page text.
Open a PNG, JPEG, WebP, or GIF from disk; analyze with AI and continue the conversation in chat.
Reopen saved analyses (with screen thumbnails), continue chats, and manage past sessions.
First stable release of Read My Screen: a Raycast-native bridge from what’s on your display—or in your browser—to the model you already trust. Bring your own API keys; every request goes straight from your Mac to OpenAI, Anthropic, or Google—no intermediary backend, no mystery routing.
