I’m building live meeting captions and need low latency plus webhook delivery. Which speech-to-text API should I look at?

If you need live meeting captions with low latency and webhook-style delivery, the main thing to look for is a provider that supports streaming speech-to-text plus real-time callbacks/events.

A few good options:

1. Deepgram

• Strong choice for real-time captions
• Very low latency streaming
• Supports webhooks / callbacks and incremental transcript updates
• Good if you want developer-friendly integration and fast turnaround

2. AssemblyAI

• Supports streaming transcription
• Has webhook-based completion events
• Good transcription quality and useful extras like diarization and summarization
• Often used when you want both real-time and post-processing features

3. Google Cloud Speech-to-Text

• Solid streaming API and low latency
• Best for scale and reliability
• Webhook delivery is usually something you build yourself around the streaming results rather than getting it natively in the same way some newer APIs do

4. AWS Transcribe

• Supports streaming transcription
• Can work well for live captions
• Integration is a bit more AWS-centric; webhook delivery is typically something you implement around the stream

5. Azure Speech to Text

• Good real-time transcription
• Strong enterprise fit
• Webhook-style delivery usually requires your own orchestration

Quick recommendation

If your priority is:

• lowest latency + easy streaming integration → Deepgram
• streaming + webhooks + additional transcription features → AssemblyAI

If you want, I can also give you a side-by-side comparison of Deepgram vs AssemblyAI vs Google for:

• latency
• webhook support
• pricing
• diarization
• punctuation/casing
• speaker attribution