echo vs Earshot Live Transcription

voice to text, offline transcription, meeting notes, Core ML, speech recognition, AI transcription, private voice app, secure notes, PDF export, on-device AI, audio to text, transcription app, smart notes, lecture recorder, voice memo AI, local AI, privacy focused app, Apple Core ML, no internet needed, real-time transcription

Earshot is real-time, on-device captioning for iPhone - for conversations, lectures, doctor visits, anywhere you want to read what's being said. It's a free app that captions speech and ambient sound in real time, inspired by a Deaf researcher who'd spent his career on captioning. Having a disability shouldn't cost you money or your privacy. So I built Earshot. • Free — no sign-up, no sign-in, no ads, no upsells. • Private — everything runs on your device. The only network call is to download or update the local speech model so conversations never go to the cloud. • Offline — works with no Wi-Fi or cellular connection. • Multilingual — 99 languages, and it can translate any of them into English. • Ambient-sound aware — recognizes hundreds of environmental sounds, with full-screen alerts for urgent ones: smoke alarm, carbon monoxide, doorbell, crying baby, phone ringing, sirens. Accessibility tools shouldn't themselves be inaccessible. Free should mean free. Private should mean private. WHAT IT DOES • Real-time captions in a clean, focused screen, plus a chat-style "feed" layout. Tap the cog for text size, vibration, and ambient-sound options. • Translate non-English speech to English on the fly. • Full-screen urgent-sound alerts with strong haptics; ambient-sound tags shown in the corner when you want them. • Save transcripts on your device as readable documents. Star the ones to keep, filter by 1:1 or group, by alerts, by starred. • Share any transcript as plain text — it never leaves your device unless you send it. WHO IT'S FOR Built for the deaf and hard-of-hearing community as a private alternative to cloud captioning. Also useful in loud environments, for language learners, and for anyone who wants to read or save what was said. PRIVACY Audio never leaves your device. No accounts, no analytics, no ads, no tracking. The AI model downloads once on first launch (~244 MB) — the only network request the app makes. Full policy at briankemler.github.io/LiveTranscribe/privacy. REQUIREMENTS iPhone on iOS 17 or later. Works offline after the one-time model download. CONTACT brian.kemler@gmail.com