Skip to content

Authentication

Pass your API key via api_key in the config:

from tarash.tarash_gateway.models import VideoGenerationConfig

config = VideoGenerationConfig(
    provider="fal",
    model="fal-ai/veo3.1/fast",
    api_key="fal-xxxxxxxxxxxxxxxx",
)

Warning

Never hardcode API keys in source files. Use environment variables or a secrets manager.

Provider API key reference

Provider Where to get it
Fal.ai fal.ai/dashboard
OpenAI platform.openai.com/api-keys
Azure OpenAI Azure Portal
Runway app.runwayml.com
Google (Gemini API) aistudio.google.com
Google (Vertex AI) GCP project + service account or Application Default Credentials
Replicate replicate.com/account
Stability AI platform.stability.ai
xAI console.x.ai
ElevenLabs elevenlabs.io
Cartesia play.cartesia.ai
Sarvam AI dashboard.sarvam.ai
Hume AI platform.hume.ai

Google models support two authentication modes

Gemini Developer API (API key): Works for both video (Veo) and image (Imagen, Gemini) models. Pass your AIza... key directly:

config = VideoGenerationConfig(
    provider="google",
    model="veo-3.0-generate-preview",
    api_key="AIza...",
)

Vertex AI (Google Cloud): Use this if you're running on GCP or need enterprise controls. Set api_key=None and supply provider_config:

config = VideoGenerationConfig(
    provider="google",
    model="veo-3.0-generate-preview",
    api_key=None,
    provider_config={
        "gcp_project": "my-gcp-project",
        "location": "us-central1",               # optional, defaults to us-central1
        "credentials_path": "/path/to/key.json", # optional, omit to use ADC
    },
)

Authentication is handled via a service account JSON key (credentials_path) or Application Default Credentials if credentials_path is omitted.