ElevenLabs STT Compatible

Speech-to-Text
on your terms

Self-hosted ASR service with multi-model support, real-time streaming, and a drop-in replacement for the ElevenLabs STT API.

Terminal
# Just change the URL — your ElevenLabs code works as-is
curl -X POST https://asr.doubtech.com/v1/speech-to-text \
  -H "xi-api-key: YOUR_KEY" \
  -F "file=@audio.mp3" \
  -F "model_id=openai/whisper-large-v3-turbo"

Everything you need for ASR

Production-ready speech recognition with the flexibility to choose your models, your infrastructure, your rules.

Multi-Model Transcription

Deploy and compare multiple ASR models side-by-side. Whisper, Qwen3, and more — pick the best model for your use case.

Real-Time Streaming

Stream audio via WebSocket for live transcription. See results appear as you speak with sub-second latency.

50+ Languages

Transcribe audio in over 50 languages with automatic language detection. From English to Cantonese, we've got you covered.

Self-Hosted & Private

Run entirely on your own infrastructure. Your audio never leaves your network. Full control over data and models.

ElevenLabs Compatible

Drop-in replacement for the ElevenLabs STT API. Change the URL and API key — your existing integration just works.

Model Manager

Deploy, monitor, and manage ASR model containers with a desktop app. One-click deployment, rebuild, and health checks.

Drop-in Replacement

ElevenLabs STT API
fully compatible

Switch from ElevenLabs to your own infrastructure by changing two lines of config. Same endpoints, same request format, same response schema. Your existing code just works.

  • Same xi-api-key authentication header
  • Identical /v1/models response format
  • Compatible WebSocket streaming protocol
  • Multipart form upload for batch transcription
  • Partial and committed transcript messages
Compatible Endpoints
GET/v1/models
POST/v1/speech-to-text
WS/v1/speech-to-text/realtime

Choose your model

Deploy multiple ASR models and compare accuracy, speed, and language support. Add new models with one click.

OpenAI

Whisper Large V3 Turbo

SpeedFast
Languages50+
Alibaba

Qwen3-ASR-1.7B

SpeedMedium
Languages30
Kyutai

Moshi STT 2.6B

SpeedAccurate
LanguagesEnglish

Get started in minutes

Three steps to your first transcription.

1

Get an API key

# Sign in and create a key at
https://asr.doubtech.com/api-keys
2

List available models

curl https://asr.doubtech.com/v1/models \
  -H "xi-api-key: YOUR_KEY"
3

Transcribe audio

curl -X POST https://asr.doubtech.com/v1/speech-to-text \
  -H "xi-api-key: YOUR_KEY" \
  -F "file=@audio.mp3" \
  -F "model_id=openai/whisper-large-v3-turbo"

Ready to transcribe?

Start using your own ASR service today. No usage limits, no data leaving your network.