Video & audio
Video and audio use the platform-native API. These jobs take
from seconds to minutes, so send them with "is_async": true and poll, or use
webhooks.
Video
Endpoint: POST /generation/v1/generate/video
curl https://api.funai.art/generation/v1/generate/video \
-H "Authorization: Bearer $FUNAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model_id": "veo3-1-fast",
"workflow": "text_to_video",
"prompt": "A drone shot of a tropical island at sunset",
"duration": 8,
"is_async": true
}'
For image-to-video, use "workflow": "image_to_video" with an image-to-video model
and pass the input image as the model's reference describes.
Audio and speech
Endpoint: POST /generation/v1/generate/audio
Text-to-speech, music and sound effects.
curl https://api.funai.art/generation/v1/generate/audio \
-H "Authorization: Bearer $FUNAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model_id": "elevenlabs-tts-multilingual-v2",
"workflow": "text_to_audio",
"prompt": "Welcome to FunAI.",
"is_async": true
}'
Voices and other model options go in custom_params. The
model reference lists them for each model.
Getting the result
- Submit: the response holds a
job_idand"status": "PROCESSING". - Poll:
GET /generation/v1/jobs/{job_id}/status. - Done: when
statusisCOMPLETED,output_urlpoints at the file.
Or subscribe to the generation.completed and generation.failed
webhooks.