Saved
Recent
Camera is off
Point at a sign, menu or page, then take one shot. The reading stays put until you scan again.
Full reading
Tap the microphone and speak a sentence.
—
Reading the picture, hearing your voice and speaking the answer all happen on this device — zero bytes. Only the translated text goes out, batched so a whole camera scan costs one request of a kilobyte or two.
Network used
last 60 minutes
Daily characters
MyMemory allowance
Voice input
Whisper base
Multilingual speech to text.
Camera text
Tesseract 7
Each language downloads its own reading data the first time you scan in it, then stays cached.
Speech output
Device voices
Reading aloud uses the voices your phone or computer already has.
Photos and audio never leave the device. Translation sends plain text to your own Cloudflare Worker and nothing else — no keys, no third party, no logging unless you add it. Libraries come from jsDelivr and the voice model from Hugging Face on first use only.