Conversational AI
<- Benchmarks

Speech-to-Text APIs

Streaming transcription services and open-source engines for conversational AI input.

Deepgram

per audio time

Website Docs
Deployment
managed API
Modalities
streaming STT, batch STT
Strengths
low-latency streaming; voice-agent ecosystem; language and domain options
Tradeoffs
cloud dependency for managed API

AssemblyAI

per audio time

Website Docs
Deployment
managed API
Modalities
streaming STT, batch STT
Strengths
developer-friendly docs; streaming APIs; speech intelligence features
Tradeoffs
managed API abstraction

Soniox

usage-based

Website Docs
Deployment
managed API
Modalities
streaming STT, multilingual STT
Strengths
multilingual coverage; streaming transcription; public benchmarks
Tradeoffs
smaller ecosystem than some larger platforms

Whisper / faster-whisper

open source

Website Docs
Deployment
self-hosted, local
Modalities
batch STT, near-realtime STT
Strengths
open-source baseline; local deployment; large community
Tradeoffs
streaming behavior and latency require engineering

What to measure next

  • word error rate
  • partial transcript stability
  • finalization latency
  • language and accent coverage
  • domain vocabulary support

V1 methodology note

This page is a curated directory, not a measured leaderboard. Scores will be added only when the methodology, scenarios, and raw assumptions are public.