Why is what I'm saying not being transcribed correctly?

We regularly test transcription providers, so that we can provide those that are the best for the language you're learning. You can switch provider in the Chat Settings.

Transcribing non-native speakers isn't easy. Depending on your pronunciation, you may occasionally see that it doesn't quite return what you intended. Frustrating at times.

Transcription models can also struggle with background noise, as well as long silences. To maximise accuracy, we recommend not speaking in a noisy environment, and pausing the recording if you are not ready to speak (if you have 'auto-record' on).

If you're having trouble being understood, you may also want to hit pause before sending your reply so you can check it before sending, and also ensure 'auto-send' is off.

In our hands-free 'call mode', we currently only have access to one transcription model, which is one of the best but not quite as accurate as the V1 model available in the standard chat modes, like 'Listen & Read'.

For standard chat modes, the V1 transcription model is recommended as it:

• Is more accurate than any other model (across the languages we've tested)
• Has an impressive ability to understand even if you switch languages half way through a sentence
• Is optimised by our team for speed
• Does not hallucinate (more in this later)
• Is quite good at filtering out background noise

The only real downside we've observed is that it will transcribe literally anything you say, which isn't ideal when you hesitate and say 'ummm...' or 'err..'. So try to avoid making noises when you're thinking 😄

We provide other models via chat settings because they can all occasionally become unstable as they try to keep up with rapidly growing demand. So if you're seeing a big delay in your speech being transcribed, or it seems stuck, you can temporarily switch to another model.

The V2 model is stable, but it will occasionally hallucinate, writing random things that you did not say. Hallucination can occur if there is noise or music in the background, or if you leave the recording on and do not speak, e.g. if you cough and don't say anything. It also has a tendency to correct your speech sometimes, which prevents the other AI models from highlighting your errors.

Feel free to experiment and see which technology works best for you! Our team is also often experimenting with new models so that we can keep offering the best available tech over time.