EchoAssist gives live, speaker-labeled captions of nearby conversation, entirely on your iPhone.
Email wen.ethan@proton.me, or open an issue on GitHub. Include your iPhone model and iOS version if something isn't working.
The speech models (about 800 MB) have to finish downloading before captions can start. Check Settings → Speech Models. If a download stopped partway, tap Resume Download.
Speaker labels need more than one voice. With a single person talking, every line is correctly labeled Speaker 1. You can rename speakers inside a saved recording.
New words appear right away in a lighter shade, then darken once the app has confirmed who said them, usually within a second or two.
Summaries use Apple Intelligence and need a device that supports it. On other devices, the summary shows the opening of the transcript instead.
No. Audio is never saved or uploaded, and transcripts stay on your device. See the privacy policy.