OpenAI Whisper API based on Node.js / Bun.sh in a Docker Container.
1.2K
This is is a Node.js OpenAI microservice using Bun.sh with Docker. With zero dependencies.
It listens to /transcribe and returns the transcription.
It includes a Docker image for the OpenAI Whisper AI.
Install bun.sh first, clone this directory and run these commands:
bun install
bun run dev
You can now navigate to http://localhost:3000 or the PORT provided, see the Usage section below.
Install bun.sh first, clone this directory and run these commands: Change the project ID to your own.
docker build --platform linux/amd64 -t gcr.io/magicbuddy-chat/whisper-docker .
docker push gcr.io/magicbuddy-chat/whisper-docker
gcloud run deploy whisper-docker \
--image gcr.io/magicbuddy-chat/whisper-docker \
--platform managed \
--region us-central1 \
--allow-unauthenticated \
--project magicbuddy-chat
You should receive a Service URL, see the Usage section below.
You can test normal HTTP by opening the /ping endpoint on the URL.
Connect to the /transcribe and send a POST request with the following body:
{
"audio": "BASE64_ENCODED_AUDIO"
}
You need to pass the OpenAI API Key as a HEADER:
Authorization: Bearer OPENAI_KEY
Or you can launch the docker image or server with OPENAI_KEY in the env:
OPENAI_KEY=YOUR_KEY_HERE bun run dev
# or
docker run -p 3000:3000 -e OPENAI_KEY=YOUR_KEY_HERE gcr.io/magicbuddy-chat/whisper-docker
# or set it as env in Cloud Run with the below command or in the Cloud Console UI
gcloud run deploy whisper-docker \
--image gcr.io/magicbuddy-chat/whisper-docker \
--platform managed \
--region us-central1 \
--allow-unauthenticated \
--project magicbuddy-chat \
--set-env-vars OPENAI_KEY=YOUR_KEY_HERE
We are using this Whisper API with MagicBuddy, a Telegram ChatGPT bot.
You can use the OpenAI Whisper Docker as a live example here:
Content type
Image
Digest
sha256:b828418e2…
Size
103.5 MB
Last updated
about 3 years ago
docker pull illyism/openai-whisper-api