Speech Transcription using WhisperX DLC¶
Production-ready Docker images for transcribing, aligning, and diarizing speech with WhisperX on AWS. Built on Amazon Linux 2023 with ongoing security patching.
Transcribe audio to text, align words to precise timestamps, label speakers, and export subtitles through an OpenAI-compatible API.
Images¶
| Platform | Image | Default Port |
|---|---|---|
| EC2 | public.ecr.aws/deep-learning-containers/whisperx:3.8.6-cu128-amzn2023 |
8000 |
| Amazon SageMaker AI | public.ecr.aws/deep-learning-containers/whisperx:3.8.6-cu128-amzn2023-sagemaker |
8080 |
All images are also available on the ECR Public Gallery. For private ECR URIs, see Image Access.
What's Included¶
In addition to WhisperX and its PyTorch/CUDA runtime stack, the images bundle:
- faster-whisper — Whisper inference engine backed by CTranslate2
- CTranslate2 — the optimized runtime that executes the Whisper models
- wav2vec2 forced alignment — per-word timestamps from phoneme-level alignment models, applied on top of the transcript
- pyannote.audio — speaker diarization pipeline, pre-baked into the image
- ffmpeg — audio decoding for any input container or codec
- FastAPI + uvicorn — the OpenAI-compatible HTTP server
API Endpoints¶
| Endpoint | Purpose |
|---|---|
POST /v1/audio/transcriptions |
Transcribe an audio file (OpenAI-compatible) |
POST /invocations |
SageMaker alias — identical behavior to /v1/audio/transcriptions |
GET /ping |
Readiness health check |
GET /v1/models |
Advertise the served model id |
The request is a multipart/form-data file upload, not a JSON body. See EC2 Deployment and
Amazon SageMaker AI Deployment for examples, and Configuration for every launch option and request field.
How We Build¶
These images are curated builds tracking the WhisperX project:
- Built from upstream releases — images track WhisperX releases, each gated by our regression test suite before publication.
- Regression-tested — validated for transcription, word alignment, and diarization on EC2 and Amazon SageMaker AI on every release. See Supported Models.
- Security-patched — continuously maintained with security patches from AWS on an Amazon Linux 2023 base.