OpenEar

Every voice deserves to be heard. Every word deserves to be understood.

Free • Open Source • Local • Private

View on GitHub Help the Cause
$0
per minute, forever
No
subscriptions
100%
local & private

What is OpenEar?

OpenEar is a free, open-source system that provides real-time AI-powered captioning and translation for church services — in 200 languages.

A small computer taps audio directly from your soundboard and uses NVIDIA's Parakeet speech recognition to transcribe sermons in real time, then translates via Meta's NLLB-200 model. Captions stream to any phone, tablet, or display on your local Wi-Fi. No app to install. No account to create. Just open a browser.

Everything runs locally. No audio leaves your building. No internet connection is required during services.

Why it's different

"We preach so that everyone can encounter God face to face. When technology can remove the glass, no one should have to pay for the window."
— Rev. Dr. Jonathan Mellette, Pastor & Developer

Roadmap

VersionFeatureStatus
v0.1Proof of concept — live transcription via browser✓ Done
v0.2Server-side mic capture from soundboard✓ Done
v0.3UI polish, quality presets, admin settings✓ Done
v0.4Real-time translation — 200 languages, any device✓ Done
v0.5One-click installer✓ Done
v0.6Mac Mini / Jetson Orin testing◐ In progress...
v0.7Testing at select churches◐ In progress...
v0.8Expanded testing & feedback○ Planned
v0.9Public release — manually deployable for any church○ Planned
v1.0Public release — automatically deployable for any church○ Planned

The story

OpenEar was built by a pastor who spent a decade as a software engineer before entering ministry. He wanted every member of his congregation to feel fully included — regardless of language, hearing ability, or what resources they had available.

Every congregation deserves accessibility — not just the ones that can pay for it.