Best speech-to-text API for voice agents
Depot Pick: Deepgram. Runner-up: AssemblyAI (then OpenAI Media (Images/Audio)). Job: Speech-to-text / transcription. Updated 2026-11-03 (newest stored date on this page).
Transcribe audio for agents.
Ready call for agents
MCP https://getagentdepot.com/api/mcp → decide {"task":"transcribe audio to text"}
HTTP GET /api/decide?jtbd=speech-to-textNo key. The reply carries the pick, runner-up, the pick's sourced price with its date, and a recheck token.
Depot Pick: Deepgram
Editorial score 8.50 / 10 (curated estimate, as of 2026-10-28). Prices below are copied from the provider's own page; units are never converted.
Blunt: openai-media was the Round-1 convenience pick before Deepgram entered the catalog. Deepgram wins production STT on $/min + streaming agent DX (Nova-3 / Flux). Keep openai-media only when you refuse a second vendor. Aligns with production-stt.
| Item | Price (USD) | Source | Price as of | Last re-checked |
|---|---|---|---|---|
| Voice Agent API Standard | $0.075 / minute | official page | 2026-10-28 | not reconfirmed |
| Nova-3 Monolingual pre-recorded speech-to-text | $0.0043 / minute | official page | 2026-11-03 | not reconfirmed |
| Nova-3 Multilingual pre-recorded speech-to-text | $0.0052 / minute | official page | 2026-11-03 | not reconfirmed |
| Whisper Large pre-recorded transcription | $0.0048 / minute | official page | 2026-11-03 | not reconfirmed |
Runner-up 1: AssemblyAI
Editorial score 8.11 / 10 (curated estimate, as of 2026-10-28). Prices below are copied from the provider's own page; units are never converted.
| Item | Price (USD) | Source | Price as of | Last re-checked |
|---|---|---|---|---|
| Voice Agent API | $4.5 / hour | official page | 2026-10-28 | not reconfirmed |
| Universal-3.5 Pro async speech-to-text | $0.21 / hour | official page | 2026-11-03 | not reconfirmed |
| Universal-2 async speech-to-text | $0.15 / hour | official page | 2026-11-03 | not reconfirmed |
Runner-up 2: OpenAI Media (Images/Audio)
Editorial score 8.15 / 10 (curated estimate, as of 2026-10-28). Prices below are copied from the provider's own page; units are never converted.
No sourced price stored (null). Depot never estimates a price.
Lowest sourced prices per minute
Same unit only, price only (not quality-adjusted). The pick is editorial and is not chosen by price.
| Provider | Item | USD / minute | Source | Price as of | Last re-checked |
|---|---|---|---|---|---|
| Cloudflare Workers AI | @cf/openai/whisper audio minute | $0.0005 | official page | 2026-10-27 | not reconfirmed |
| DeepgramDepot Pick | Nova-3 Monolingual pre-recorded speech-to-text | $0.0043 | official page | 2026-11-03 | not reconfirmed |
Full list: /cheapest/speech-to-text
Questions
- What does Agent Depot pick for Speech-to-text / transcription?
- Deepgram is the stored Depot Pick for the job speech-to-text. Blunt: openai-media was the Round-1 convenience pick before Deepgram entered the catalog. Deepgram wins production STT on $/min + streaming agent DX (Nova-3 / Flux). Keep openai-media only when you refuse a second vendor. Aligns with production-stt.
- What is the runner-up?
- AssemblyAI, then OpenAI Media (Images/Audio). The ranking is editorial (scores as of 2026-10-28) and is not chosen by price.
- When should I not use Deepgram?
- Already all-in on OpenAI and Whisper quality/latency is fine — openai-media; need richest audio-intelligence features (summaries/PII) — AssemblyAI; ultra-cheap batch Whisper on Groq LPU — groq.
- How much does Deepgram cost?
- Voice Agent API Standard: $0.075 per minute, copied from https://deepgram.com/pricing (price as of 2026-10-28). More rows are in the table on this page.
Picks are editorial and dated; community reports never change a pick or a score. Method: /methodology · Corrections: /corrections · Compare: Deepgram vs AssemblyAI