I tried different configurations for Gladia transcriber, based on the API, but no input ever got transcriber. Call ids (examples) : - 19f11459-2992-4ec1-bc43-c00995f7a487 - bca57c39-4915-48c3-86cd-6b6a13d07fc4 - 96db6f00-853f-44c8-a0e8-b802190c057f Each time I've talked and waited a few seconds. As there was no answer, I've ended the call and check the transcription the call : there was nothing transcribed from what I said. It seems like gladia transcriber instantiation didn't work? Here are the different set-ups I've tried : (without model) "transcriber": { "languageBehaviour": "manual", "language": "fr", "provider": "gladia", "transcriptionHint": "Florence Wéra médecin généraliste Genappe", "confidenceThreshold": 0.1, } (basic, as before) "transcriber": { "model": "fast", "language": "fr", "provider": "gladia" } (mix of both) "transcriber": { "model": "fast", "languageBehaviour": "manual", "language": "fr", "provider": "gladia", "transcriptionHint": "Florence Wéra médecin généraliste Genappe", "confidenceThreshold": 0.1, } ...

Hi there! 👋 I see you're having trouble with the Gladia transcriber not working despite trying different configurations. Unfortunately, I don't have specific information about your particular call IDs or the exact definitions of some parameters like

languageBehaviour

,

transcriptionHint

, and

confidenceThreshold

for Gladia. However, I can help troubleshoot the general issue with the Gladia transcriber not working. Based on the information I found, here are some potential reasons and solutions for your Gladia transcriber issue: ## Potential Network Issues One common cause could be network-related problems. According to [Discord discussions](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1340235729204285510), users have experienced similar issues with error messages like:

Transcriber failed (reason: Received network error or non-101 status code.)

This could indicate IP blocking or connectivity issues between Vapi and Gladia. ## Configuration Issues Your configuration looks generally correct, but there might be subtle issues: 1. **Language Settings**: Make sure the language setting (

"fr"

) is properly supported by Gladia. According to [troubleshooting information](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1328595920899674152), incorrect language settings can cause transcription failures. 2. **SDK Version**: Ensure you're using the latest version of the Gladia SDK, as mentioned in [troubleshooting guides](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1328595920899674152). 3. **Endpointing Settings**: Some users have resolved transcription issues by adjusting endpointing settings. You might want to try adding:

json
"endpointing": 300

to your transcriber configuration. ## Alternative Approaches If Gladia continues to fail, consider trying a different transcriber provider. Multiple users have reported success after switching:

@Sahil could you take a look at this one ? 🤔

🔵 15:34:20:532 Gladia

user

Websocket Closed. Code: 4400, Reason: 'prosody' is not a valid parameter. Do you mean 'frames'? Issue ^^

Hello @Sahil ! Prosody ? Never used this parameter, as you can see from my examples above 🤔 This must be an issue on Vapi side I believe. To make sure, I've just created a brand new assistant via the VAPI UI, with Gladia as transcriber - it doesn't work, nothing is sent to Gladia I believe. - New assistant id : d043da95-217e-4e13-8760-80d2ba2573c7 - New call id : efcaa4c4-cf0e-4da9-b536-e9c796adf8e8 Assistant transcriber object (used GET /assistant/:id in the VAPI API): "transcriber": { "model": "fast", "prosody": false, "language": "fr", "provider": "gladia", "audioEnhancer": false, "confidenceThreshold": 0.1 },

Yeah, It is from the our side. I sent the details to the team and will inform you once it is resolved.

Thanks @Sahil . Indeed, I'm using another transcriber in the meantime. Btw, Gladia's new Solaria model allows to pass an array of languages as language parameter (instead of "multi" that would replace all of them, you can choose a subset of all languages). Can you pass this to the team as well ?

Sure.

@Sahil any news on your side ?

Sorry, but we are pretty swamped right now. So, there hasn’t been any progress on it.

Ok, thanks for the transparency. I believe gladia could be a game changer for multilingual though, I will try to get people warm for it so you'd have even more good reason to integrate it !

You can raise an feature request for it over here: https://roadmap.vapi.ai

@Shubham Bajaj @Kings_big💫 it's there already : https://roadmap.vapi.ai/feature-requests/p/add-new-gladia-transcriber-model-solaria

We'd love to have this too

@Kings_big💫 @Shubham Bajaj @Sahil it feels like it's been picked up by many EU-based devs I've talked to. Now top of requests lists 😇 https://roadmap.vapi.ai/feature-requests/p/add-new-gladia-transcriber-model-solaria Would be very happy if you could share progress once there is some 🙏

Thank you!

@Sahil isn't it supposed to show "under review" or something on the feature request if this has been prioritized somehow by the team ? 🤔 https://roadmap.vapi.ai/feature-requests/p/add-new-gladia-transcriber-model-solaria