Plainly

Roughly one in five people struggle with complex text — dyslexia, low literacy, aphasia, a non-native language, a cognitive disability. The everyday written world — medical forms, official letters, dense signage, a menu in a language you don't read — quietly shuts them out. Tools exist to translate or read text aloud, but they rarely make dense information genuinely understandable, and they usually send your most private documents to the cloud to do it.

Plainly turns your phone's camera into a plain-language lens. Point it at any text and it comes back rewritten in clear words, in your language, and read aloud — and you can ask it questions, like 'what is this form actually asking me for?'. It's a sibling to Dot Go: the same accessibility DNA aimed at a fresh problem — making the written world legible to anyone.

The build is native for Android and designed to run the whole loop on-device on flagship phones using Google's edge stack, so nothing leaves the phone. Point: ML Kit extracts the text locally in about 50ms. Ask: Android's native Speech Recognizer transcribes your question. Reason: Gemini Nano, via AICore, runs entirely on the phone's NPU. Listen: Android TTS reads the answer aloud. The result is a complete Point → Ask → Listen loop with zero network latency and full on-device privacy.

Built with Gemini 3 in Google Antigravity; in active development. The demo here runs the same flow through a cloud pipeline so it plays flawlessly in the emulator, while the on-device build is ready — a working proof that a genuinely private, low-latency accessibility assistant can now live entirely on the device in your hand.