• Open source
  • Free for everyone
  • Funded by donations

Fix the sounds that hold you back. Free, for everyone.

Glosso listens to you speak, scores every sound down to the phoneme, and tells you exactly which one to fix. It is open source, it has no paywall and no ads, and it runs on donations from the people who use it.

Android 8.0 and newer. iOS is in progress. No account needed.

A real Glosso result: the sentence "I think it is the third thing" scored 89 out of 100, and the sounds panel flags /θ/ at 33 while /a/ and /d/ sit at 100.
A real result: I think it is the third thing. scored 89/100. One sound slipped, and Glosso names it.
  • 0 €to use every lesson and feature. There is no paid tier.
  • AGPLthe licence. App, engine and server are all public.
  • 0ads, data sales or third-party trackers. Ever.
  • 100%funded by voluntary donations from learners.
How it works

Speak. Measure. Improve.

One short loop, repeated until it sticks.

  1. Speak

    Read one short phrase aloud. Glosso records it and scores the audio, on your phone or on the server.

  2. Measure

    A speech model turns the audio into the sounds it heard. Each one is aligned against the reference and scored, not just the sentence.

  3. Improve

    Glosso names the sound that slipped, explains the fix in plain language, and builds your next drill from your own error.

Glosso Today screen: today's lesson "A Call Home" with its "Greet a family member on the phone and agree a time to be home." description, the day streak, and the start of the "Sounds to work on" section.
Your day, planned. Today's lesson, your streak, and the sounds still waiting to be worked on.
Glosso practice screen: the sentence "The path goes past the north gate." with its IPA transcription and the path / pass sound pair, ready to record.
Say it, then see it. The sentence to record, its IPA, and the exact sounds it targets.

No speech model is infallible. When the audio is too quiet or too noisy to judge, Glosso says the attempt could not be evaluated instead of inventing a low score. See the FAQ for the honest limits.

Open source

Every line of code is public.

Glosso is free software under the GNU AGPL. You can read how a score is computed, check what the app sends, and run the whole thing yourself.

  • Read the scoring. The phonetic engine that judges your voice is the same code on the phone and on the server, and you can audit every step.
  • Run your own server. The server ships with a Docker setup. Point the app at it in Settings > Advanced > Server.
  • No Google required. An F-Droid build leaves out every Google library and starts with usage statistics off.
  • Make it better. Bug reports, fixes, translations and new languages are all welcome.
git clone https://github.com/IgnacioLD/glosso
engine/
phonetic engine and scoring
server/
API, content and assets
shared/
the app, shared by Android and iOS
androidApp/
Android shell, Play and F-Droid builds
iosApp/
iOS app, in progress
deploy/
self-hosting: Docker and Compose

AGPL-3.0-or-later

Get Glosso

Start practising today.

Install it from Google Play, or build it yourself from the source.

Android 8.0 and newer. Scored on the server by default, or on your phone with the optional model. No account, no ads, no third-party trackers.