Hey, is anyone else seeing this behavior? Vapi emits a clean partial transcript (e.g. "Hey."), then suddenly sends a final transcript like "A.", which overwrites the partial in the UI. In my code, I’m following standard handling – on isFinal, I replace the partial with the final. But in this case, it feels like the final result is either truncated or incorrect, leading to bad UX (just showing "A."). I’m wondering if this is a known issue, maybe something with the ASR backend or a race condition? Or if it’s just me and I need to patch something client-side. Would love to know if others are seeing this or have a fix. https://cdn.discordapp.com/attachments/1369082752493686794/1369082752661717133/Screenshot_2025-05-05_at_11.52.15.png?ex=681a912c&is=68193fac&hm=e8ee8f14ed36378ddd5235ad7041e3915c869cd4f1b9dfb4995eaffcac59abe6&

This appears to be an endpointing issue with Vapi's transcription service. Try increasing the

waitSeconds

parameter to 0.8 and consider implementing a custom endpointing solution:

javascript
// Configure your assistant with longer wait time
const assistant = new Assistant({
  // other config...
  waitSeconds: 0.8,
  // Optional: custom endpointing
  endpointing: {
    type: 'custom',
    value: '300ms'
  }
});

Several users have reported similar issues, particularly with Deepgram. If the problem persists, try switching to another transcriber like Azure or Whisper. Sources: - [Discord conversation about endpointing issues](https://discord.com/channels/1211482211119796234/1343636482874933340) - [Discord conversation about transcriber issues](https://discord.com/channels/1211482211119796234/1334024569861963788)

Hey! To help track down this issue, could you share: - The call ID - When exactly this happened (the timestamp) - What response you expected to get - What response you actually got instead This would really help us figure out what went wrong!