On Azure docs, their STT supports language detection. Is there any plan for Vapi to support this as well?

Hi there, Thank you for your message. Our team is currently out of the office. We operate Monday through Friday, from 9:00 AM to 8:00 PM Pacific Standard Time (PST). We’ll get back to you as soon as possible during our normal business hours. If your message is urgent, please mark it accordingly or include “URGENT” in the subject line, and we’ll do our best to respond promptly. Warm regards, Vapi Customer Support Team

Currently, Vapi does not support automatic language detection with Azure STT. If you need auto-detection, use a provider like Deepgram with the

"language": "multi"

setting:

json
"transcriber": {
  "provider": "deepgram",
  "model": "nova-2",
  "language": "multi"
}

This enables seamless multilingual support in Vapi. There is no public information about future support for Azure's language detection in Vapi at this time. Source: - [Vapi Multilingual Configuration Guide](https://docs.vapi.ai/customization/multilingual)

Deepgram is much worse for our use case. Azure has language detection already, is there plan to support that soon or should I host it as a custom transcriber?

Hi Bathroom King, Currently, when using Azure as a transcriber, you need to specify the language explicitly:

{  "transcriber": {    "provider": "azure",    "language": "pt-BR"  // Language must be specified  }}

There's no mention of automatic language detection features or plans in the current documentation. For multilingual use cases, you would need to: 1. Know the expected language beforehand 2. Configure the transcriber with the specific language code 3. Use different assistant configurations for different languages

I understand. But what I am saying is that Azure has support for language detection and I'm wondering if Vapi plans to add that? Otherwise I will need to self-host Azure just so I can enable language detection myself.

We previously had language detection in the past, but it is no longer supported on our platform. Currently, there are no plans to add it back to our platform so self-hosting might be the way to go.

Why not? Did it not perform well or was difficult to support for some reason?

We are not able to comment on this at the moment