Point your phone camera at the computer screen and record a video while scrolling through the questions. Keep each question steady on screen ~1s. Angled or off-axis shots are fine — the vision model corrects for perspective.
Stop recording, come back here, upload the video below.
Frames are extracted on-device and duplicates dropped, then batched to Google Gemini 2.0 Flash which reads the questions and answers straight from the images. DeepSeek can optionally double-check the answers.
Upload phone recording
MP4 / MOV · up to ~5 min · tap to choose
idle
API keys
Stored only in this browser (localStorage). Google Gemini reads the frames (vision). DeepSeek is optional — when enabled it refines the answers as a second pass.