vapi websdk - live transcript events suddenly not ...
# support
j
something change or break in webSDK?
v
Hi there! πŸ‘‹ I understand you're experiencing an issue with live transcript events in the Vapi WebSDK suddenly not working. While I can't definitively confirm if something specific changed or broke in the WebSDK, I can share some relevant information that might help. Several users have reported similar issues with live transcript events in the Vapi WebSDK. According to [user reports in Discord](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1334660289668583514), there have been instances where live transcript events suddenly stopped working. Here are some recent changes that could potentially affect the functionality: - Azure Speech Transcriber support was added, which might affect transcription functionality ([Changelog](https://docs.vapi.ai/changelog/2024/12/13)) - Updates to various nodes including
Say
,
Edge
,
Gather
, and
Workflow
nodes with increased character limits and new properties ([Changelog](https://docs.vapi.ai/changelog/2025/1/29)) - Introduction of Hooks in workflows that could impact event handling ([Changelog](https://docs.vapi.ai/changelog/2025/2/4)) For troubleshooting, you might want to try: 1. **Check your webhook configuration**: Ensure your webhook is correctly configured with the proper URL and settings ([Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1336399188665040968)) 2. **Try a different transcriber**: Some users resolved similar issues by switching from Deepgram to Whisper or Azure as a transcriber ([Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1343612796717371393)) 3. **Consider using webhooks as a workaround**: One user found success by "using the webhook to response then it is correct, rather than using Vapi module" ([Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1334661059298066514))
4. **Check your code implementation**: Make sure you're properly handling the transcript events. The WebSDK uses
message
events to send all messages, including
transcript
messages, which need to be filtered by
type
and
transcriptType
([Documentation](https://docs.vapi.ai/examples/pizza-website)) Is there any specific part of your implementation you'd like me to help troubleshoot further? Sources: - [Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1334660289668583514) - [Changelog](https://docs.vapi.ai/changelog/2024/12/13) - [Changelog](https://docs.vapi.ai/changelog/2025/1/29) - [Changelog](https://docs.vapi.ai/changelog/2025/2/4) - [Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1336399188665040968) - [Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1343612796717371393) - [Discord thread](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1334661059298066514) - [Documentation](https://docs.vapi.ai/examples/pizza-website)
t
seeing the same thing
@Shubham Bajaj this may be an error, in case you are not aware πŸ™‡
k
There is no such error reported as of now. Could you guys share the steps to reproduce it, and I'll take a look on my end?
j
No errors in console or events loggable. Transcription events are not coming in at all during call
Just check events being sent during a call. Using 11labs voice and deepgram transcription with β€˜en’ language
@Shubham Bajaj any updates on this?
seems to be working again now. was issue on your end correct?
k
The issue was on our end, and it has been fixed.