Soniox

by Soniox

Real-time multilingual ASR API with low-latency streaming and code-switching support.

TL;DR

Real-time multilingual ASR API with low-latency streaming and code-switching support.

Best for voice agents and call-center products that need low-latency real-time multilingual STT. Pricing: per-minute (see Soniox pricing page).

Category
Transcription APIs
License
Stars
Last push
Pricing
per-minute (see Soniox pricing page)
Platforms
API

What it is

Soniox is an independent speech-recognition API vendor focused on real-time streaming. Notable for code-switched multilingual recognition (multiple languages in one utterance), speaker diarization, custom vocabulary, and low end-to-end latency that targets agent-assist and live captioning. Available via WebSocket and REST. Pricing is per-minute with volume tiers; HIPAA-eligibility and enterprise terms via direct contract. Best fit: voice agents and call-center products that need low-latency real-time multilingual stt. Caveats: smaller vendor than hyperscalers; slas and roadmap depend on direct sales contact. Pricing as listed: per-minute (see Soniox pricing page). Feature flags from vendor docs: speaker diarization, word-level timestamps, streaming. Directory tags: commercial-api. Last vendor-page check: 2026-05-12.

Best for: Voice agents and call-center products that need low-latency real-time multilingual STT.
Watch out for: Smaller vendor than hyperscalers; SLAs and roadmap depend on direct sales contact.

Install / use

Soniox WebSocket: wss://api.soniox.com/transcribe-websocket

Features

Speaker diarizationYes
Word-level timestampsYes
Streaming / real-timeYes
Languages supportedNone
HIPAA eligibleNo

Soniox vs Whipscribe

FeatureSonioxWhipscribe
CategoryTranscription APIsTranscription APIs
Pricingper-minute (see Soniox pricing page)free beta
Speaker diarizationYesYes
Word timestampsYesYes
StreamingYesNo
Languages99
PlatformsAPIWeb, API, MCP

Alternatives to Soniox

Whipscribe is a managed faster-whisper + whisperX service. If you want transcripts without running infrastructure, paste a URL or drop a file in the form below — you'll have a transcript in seconds.