EyeStill: AI Subtitles & Translation for YouTube

Subtitles at the speaker's lips that follow whoever is talking, translated in context. Plus AI answers at one point of focus.

EyeStill puts YouTube subtitles right at the speaker's lips, so you read the captions and watch the face in one glance instead of bouncing between the face and the bottom of the screen. It follows whoever is talking, translates subtitles with AI in context, and reads ChatGPT, Claude and Gemini answers at a single point of focus. SMART SUBTITLES FOR YOUTUBE • Captions at the speaker's lips: each word appears the moment it is said, timed word by word from YouTube's own captions. • Follows the active speaker in interviews and podcasts, and moves with every scene cut. Face detection runs on your device and ignores things that only look like a face. • Made for 1.5× and 2× playback: fast speech stays readable. • Stays readable on bright scenes: white text with a heavier outline and shadow. • No repeated words from YouTube's rolling captions, and missing punctuation in auto-generated captions is restored. • Pause the video and the drum stays on screen: see the whole sentence, hover the drum to reach the settings, click any word to jump the video to it. • Make the text bigger or smaller with A− / A+. • 3D drum: the next words curve in from the right, the read ones out to the left. A flat line is one click away. AI SUBTITLE TRANSLATION • Translate YouTube subtitles into 53 languages: pick yours from a list you can search, right on the player. • An AI model reads the whole talk, so terms stay consistent, phrasing is natural and filler words are left out. • Each translated line appears at the original's timing, right at the speaker's lips: keep the real voice, read the meaning. • On pause, the original sentence shows under the translation. OTHER VIDEOS • Drop your own .srt / .vtt file onto a video, or allow EyeStill on another site from the menu. YouTube is where it works best. AI ANSWERS AT ONE POINT OF FOCUS • In ChatGPT, Claude, Gemini and DeepSeek a reading bar sits under the latest answer, with its reading time, and starts reading while the model is still writing. • Visible reasoning, citations and "copy code" labels are left out. • Alt+Q reads the latest answer, or any text you select on any page. WHO IT'S FOR • Watching lectures, podcasts and interviews at 2×. • Fluent, but not native: fast speech, accents and slang become easy to follow. • Watching videos in a foreign language with the original voice instead of dubbing. BOOKS • Alt+R opens a video's transcript in the free EyeStill speed reader at eyestill.com. It reads EPUB, FB2 and TXT too. PRIVATE BY DESIGN • Face tracking runs in your browser; the video never leaves your device. • Text leaves your device only for features you turn on: subtitle translation, and your own AI provider if you connect one. • No account, no ads. Free during the open beta. Privacy policy: https://eyestill.com/privacy · Help: https://eyestill.com/support