VoxatlasBest Speech-to-Text APIs for Voice Agents (2026)
Deepgram, AssemblyAI, Speechmatics, Gladia and Soniox compared on published real-time transcription pricing.
| Tool | Starting price | Best for |
|---|---|---|
| Nova-3 streaming $0.0048/min ($0.29/hr); $200 free credit | Low-latency speech-to-text for voice agents | |
| Streaming from $0.15/hr; $50 free credits | Accurate, inexpensive transcription | |
| From $0.24/hr pay as you go (as low as ~$0.12/hr with discounts) | Multilingual transcription with language switching | |
| ~$0.10/hr async; ~$0.12/hr real-time | Very low-cost speech-to-text | |
| Async $0.61/hr; real-time $0.75/hr | Real-time multilingual transcription |
Vapi vs Retell AI →Vapi vs Bland AI →Retell AI vs Bland AI →Synthflow vs Vapi →Smith.ai vs Dialzara →Goodcall vs Rosie →ElevenLabs Agents vs Vapi →Deepgram vs AssemblyAI →Dialzara vs Rosie →
Prices from vendors' official pricing pages, checked October 14, 2026. Not sponsored; order isn't a ranking.