We are prototyping a Voice-Bot for doctor in vienna. They have a variety of languages - central europe basically. Problem is VAPI cannot understand sliglhlty exotic names properly. Unfortunately Spelling does not work either. How can we fix muiltilingual transcription and increase data acuracy. Especially with a mix of central european languages, the transcriber does not work good enough. Maybe it´s my settings? If someone have solved this issue, I´m happy to connect

@MAX_HIRSCHL#1177 I'm glad to hear you're working on a Voice-Bot project! Multilingual transcription can indeed be challenging, especially with names and terms that are unique or less common. Here are a few suggestions that might help improve transcription accuracy for your project: 1. **Language Selection**: Ensure that the transcriber is set to the correct language or dialect that is being spoken. VAPI supports multiple languages, so selecting the right one is crucial for accuracy. 2. **Custom Vocabulary**: If VAPI supports it, you can add a custom vocabulary to the transcriber. This would involve adding the names and specific terms that are commonly used by the doctors in Vienna to the system so that it can recognize them more accurately. 3. **Acoustic Models**: Use or train acoustic models that are tailored to the accents and speech patterns of Central European languages. This can greatly improve the transcriber's ability to understand slightly exotic names. 4. **Audio Quality**: Ensure that the audio input is of high quality. Background noise and poor microphone quality can significantly affect transcription accuracy. 5. **Post-Processing**: Implement a post-processing step where the transcriptions are checked and corrected manually or with an automated system that can cross-reference terms against a database of known names and terms. 6. **Feedback Loop**: Use a feedback loop to continuously improve the system. As users interact with the Voice-Bot, collect the data and use it to refine the transcriber's performance over time. If you're still facing issues with the settings or need more specific guidance, please provide additional details about your current configuration, and I'll do my best to assist you further. If you're looking for someone who has solved a similar issue, I can help connect you with support for more specialized assistance. Would you like me to do that?

The best way I found for accuracy is having the user spell their name letter by letter

Ok, @Jotham thank you so much! I tried spelling, but the VAPI is interrupting a lot during the process, do you have a tip on how to change settings to make spelling smoother?

@MAX_HIRSCHL#1177 can you share the call id where Vapi is interrupting a lot during the call.

@Shubham Bajaj hi! Here is the DEMO Link: https://vapi.ai?demo=true&shareKey=e88b0ee3-4d90-43ed-a907-32dd1531c0ff&assistantId=93bfc4cb-06ff-49e9-9e1f-13a0d7405cfd

i can't help you with this, I require the call id please.

858218b5-1e7a-415d-afb3-9142f4bade47

try this assistant config: https://dump.sahilsuman.me/assistant1.json