Real-Time Translation API

Real-Time Translation API for Live Streams

StreamTranslate provides real-time translated subtitles via browser source URL — essentially an API for live translation that integrates with any streaming software.

Try Free — 60 Seconds to Live Subtitles

Last updated: April 2026

For Developers

StreamTranslate's translation pipeline processes audio in real time: WebSocket audio streaming to Deepgram for speech-to-text, multi-provider translation engines for accuracy, and Socket.io for sub-2-second subtitle delivery. The output is a browser source URL that renders translated captions — no custom integration code required.

Technical Architecture

StreamTranslate's real-time translation pipeline is designed for low latency and high reliability on live streams:

Layer 1
Audio Capture
Browser-based audio capture via Web Audio API. Microphone audio is streamed as raw PCM data over WebSocket to the server.
Layer 2
Speech-to-Text
Audio is forwarded to Deepgram's real-time STT API. Deepgram returns partial and final transcripts with low latency and high accuracy across 30+ languages.
Layer 3
Translation
Final transcripts are routed through multiple translation providers (selected per language pair for optimal quality). Translation happens server-side with minimal latency.
Layer 4
Delivery
Translated text is pushed to connected clients via Socket.io. The browser source URL renders captions as a styled overlay — ready for OBS, Streamlabs, or any streaming software.

Integration: How to Use StreamTranslate

StreamTranslate is designed to work without custom API integration code. The browser source URL is the integration point:

// 1. Go to streamtranslate.live/control
// 2. Configure source + display language
// 3. Copy the browser source URL
// 4. Add as Browser Source in OBS (1920x1080)

Browser Source URL format:
https://streamtranslate.live/overlay?session=YOUR_SESSION_ID&lang=es

The browser source URL connects via Socket.io and renders translated captions in real time. It works with any software that supports browser sources: OBS, Streamlabs, XSplit, vMix, and more.

Technical Specifications

SpecDetails
Audio inputWebSocket (raw PCM, browser Web Audio API)
STT engineDeepgram (real-time streaming)
TranslationMulti-provider (optimized per language pair)
DeliverySocket.io (real-time push)
End-to-end latencyUnder 2 seconds
Languages28 supported languages
Output formatBrowser source URL (HTML/CSS overlay)
Compatible softwareOBS, Streamlabs, XSplit, vMix, any browser source

Live Streaming Market Stats

The technical demand for real-time translation infrastructure is driven by massive global streaming audiences.

75%
of Twitch viewers are outside the US
8.5B
hours watched on Twitch per quarter
$345B
live streaming market by 2030
1 in 4
internet users watch live streams weekly

Pricing for Developers & Integrators

StreamTranslate uses simple flat pricing — no per-API-call billing, no usage metering:

  • Stream Pass: $9.99 — 24-hour access for testing or one-off events
  • Starter: $14.99/mo — for regular integration use
  • Pro: $34.99/mo — dual-language output, advanced customization
  • Unlimited: $79.99/mo — all 30+ languages, maximum throughput

Frequently Asked Questions

Does StreamTranslate have an API for real-time translation?

StreamTranslate provides real-time translated subtitles via a browser source URL that integrates with any streaming software. Under the hood, it uses WebSocket audio streaming to Deepgram for speech-to-text, multi-provider translation engines, and Socket.io for real-time subtitle delivery.

How does StreamTranslate's translation pipeline work technically?

Audio is streamed via WebSocket to Deepgram's speech-to-text engine. The transcript is then routed through multiple translation providers for optimal quality. Translated text is delivered back to the client via Socket.io in real time — under 2-second end-to-end latency.

Can I integrate StreamTranslate with custom streaming setups?

Yes. StreamTranslate outputs a standard browser source URL that works with OBS, Streamlabs, XSplit, vMix, and any software that supports browser sources. The URL renders translated captions as an overlay — no custom API integration code required.

What speech-to-text engine does StreamTranslate use?

StreamTranslate uses Deepgram for speech-to-text transcription, chosen for its low latency, high accuracy on live audio, and support for 30+ languages. Translation is handled by multiple providers to ensure quality across language pairs.

Add real-time translation to your streaming setup

Start Free — No API Key Needed

More Alternatives

Maestra Alternative Dababel Alternative LocalVocal Alternative StreamCC Alternative Wordly Alternative Kudo Alternative