The writing is on the photo, not in a note
Passwords, amounts, gate numbers, and serial codes live as pixels. Gallery search can’t reliably read them for you.
You don’t need to type everything yourself. When you send a photo or screenshot, Tavi looks at it the way a helpful person would: it reads the writing on the image, and it notices what kind of thing it is (receipt, boarding pass, whiteboard, sticker…). Later you just ask.
For anyone who photographs passwords, receipts, passes, and boards — then can’t find them in the gallery
Talk to Tavi on WhatsAppYour phone stores thousands of pictures. It does not understand “the Wi‑Fi sticker” or “the Dubai boarding pass.” You do — but only if you can find the right thumbnail.
Passwords, amounts, gate numbers, and serial codes live as pixels. Gallery search can’t reliably read them for you.
UPI proofs, QR codes, and chat snippets blur together. You remember “Nature’s Basket ₹1,240,” not which Tuesday you snapped it.
Guests want Wi‑Fi. The driver is outside. You’re pinch-zooming through months of photos under pressure.
Think of it as: “Tavi looks carefully, then files it where you can ask in normal language.”
Photo from Camera, a screenshot, or a forward. You can add a short line if you want (“office wifi”) — helpful, but not required.
If the image has text — a sticker, receipt, boarding pass, whiteboard, sign — Tavi tries to read those words and numbers so they become searchable later.
Beyond the letters, it picks up simple sense of the scene: this looks like a receipt, a boarding pass, a place photo, a whiteboard, a product label, and so on. That helps when you ask “boarding pass” instead of quoting exact text.
“Wifi password kya hai?”, “boarding pass to Dubai”, “grocery receipt.” Tavi finds the photo you saved — and can send the original image back in chat.
Same photo → Tavi understands enough to find it when you ask.
You remember the meaning. Tavi already looked at the picture for you.
Hunt the router sticker in Photos.
Ask in Hinglish; get the answer from the photo you saved.
Scroll for the boarding pass under time pressure.
Ask “boarding pass to Dubai” and get the image back.
Guess which screenshot is which.
Ask by shop or amount; forward the right one.
1) It keeps the original picture so you can get the same image back later.
2) It reads writing that appears in the picture — names, amounts, dates, passwords, gate numbers, addresses — when the text is clear enough to read.
3) It notes what kind of photo it seems to be (for example receipt, boarding pass, whiteboard, sticker, place photo) so questions like “boarding pass” or “whiteboard from class” have a better chance of hitting the right memory.
You still talk in normal language. You don’t need to know words like “OCR” or “AI vision.” Those are just how Tavi looks carefully at what you sent.
Clear, well-lit photos work better than blurry ones. If the text is tiny, a tighter crop helps.
A one-line comment after saving (“office wifi”, “Dad’s medicine strip”) is optional insurance — useful when the picture is messy.
Tavi does not name people from faces. Stickers (the WhatsApp sticker kind) aren’t saved as memories. Videos aren’t supported yet.
There’s a size limit per photo depending on your plan. Very blurry or stylised text may not read perfectly — you can still add a short typed comment.
Don’t send government IDs, card numbers, or OTPs you’d regret keeping in a notes app.
Try one real thing from your day. Send it to Tavi. Ask for it later.
Talk to Tavi