Getting Started

Install the CLI

The Fibre CLI is available on npm. Install it globally so you can run fibre from anywhere.

npm install -g @fibre-dev/cli

Authenticate

Run fibre setup to connect the CLI to your Fibre account. You will be prompted for your API key, which you can create from the dashboard or with fibre auth create-key.

fibre setup

Your credentials are stored locally and used for all subsequent commands. Run fibre auth status at any time to verify which account is active.

Deploy your first app

Create a file called fibre_app.py that defines your app's environment, GPU, and exposed port.

import fibre

image = (
    fibre.Image("nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04")
    .apt("ffmpeg")
    .pip("faster-whisper==1.1.0", "fastapi[standard]")
)

app = fibre.App("whisper-api", image=image, gpu="T4", port=8000)

Then run the deploy command to build the image and launch your sandbox:

fibre deploy fibre_app.py --name whisper-api --gpu T4 --port 8000

Fibre packages your configuration, builds the image with the specified dependencies, provisions a GPU sandbox, and returns a live URL, all in a single command.

Your app is live

Once deployment finishes, your app is available at:

https://whisper-api--yourorg.fibre.run

The sandbox starts on the first request and scales to zero when idle, so you only pay for the compute you actually use. There are no minimum fees or reserved capacity requirements.