ScreenAI vs Yoink

ScreenAI brings screen capture, annotation, and AI questions about captured content into one desktop workflow. It runs on Linux, macOS, and Windows, and can capture regions, full screens, windows, freeform selections, or scrolling content. Timed delay and a precision magnifier are available during capture. Annotation tools include drawing, arrows, shapes, text, step numbering, highlighting, blur, crop, and eraser. ScreenAI extracts text from captures and can mask emails, numbers, and IP addresses. AI options include Claude, GPT-4o, Gemini, Mistral, Grok, and Ollama, with model switching, Best Answer mode, web search, and persistent memory. Captures are saved and OCR-indexed automatically; users can chain, batch process, pin, or send them to AI together. Exports include PNG, JPEG, PDF, and animated GIF. The site describes the app as open source under AGPL and says data stays between the user and their AI, without a middleman server or telemetry. The free plan allows 10 AI requests per day; Premium costs $7 a month and includes unlimited requests.

Yoink has no write-up on Specifiction yet; its figures are above.