Pixverse (via Fal.ai)¶

Pixverse v5 and v5.5 models with transition, effects, swap, and lipsync support.

Quick Example¶

from tarash.tarash_gateway import generate_video
from tarash.tarash_gateway.models import VideoGenerationConfig, VideoGenerationRequest

config = VideoGenerationConfig(
    provider="fal",
    model="fal-ai/pixverse/v5",
    api_key="YOUR_FAL_KEY",
)

request = VideoGenerationRequest(
    prompt="A vibrant coral reef teeming with colorful fish",
    duration_seconds=5,
)

response = generate_video(config, request)
print(response.video)

Supported Models¶

Model prefix	Duration	Image-to-Video	Notes
`fal-ai/pixverse/v5`	`5s`, `8s`, `10s`	✅	Transition, effects, swap
`fal-ai/pixverse/v5.5`	`5s`, `8s`, `10s`	✅	Same API as v5
`fal-ai/pixverse/swap`	—	✅	Swap variant
`fal-ai/pixverse/lipsync`	—	—	Lipsync (video + audio/TTS)

Parameters¶

Parameter	Required	Notes
`prompt`	✅
`duration_seconds`	—	`5`, `8`, or `10`
`image_list` (reference)	—	Image-to-video

Lipsync¶

The fal-ai/pixverse/lipsync model syncs lip movements in a video to audio. You can provide audio directly via audio_url, or use built-in TTS by supplying text and voice_id.

With Audio URL¶

from tarash.tarash_gateway import generate_video
from tarash.tarash_gateway.models import VideoGenerationConfig, VideoGenerationRequest

config = VideoGenerationConfig(
    provider="fal",
    model="fal-ai/pixverse/lipsync",
    api_key="YOUR_FAL_KEY",
)

request = VideoGenerationRequest(
    video_url="https://example.com/speaker.mp4",
    audio_url="https://example.com/speech.mp3",
)

response = generate_video(config, request)
print(response.video)

With TTS (text + voice)¶

config = VideoGenerationConfig(
    provider="fal",
    model="fal-ai/pixverse/lipsync",
    api_key="YOUR_FAL_KEY",
)

request = VideoGenerationRequest(
    video_url="https://example.com/speaker.mp4",
    text="Hello, welcome to our demo!",
    voice_id="Emily",
)

response = generate_video(config, request)
print(response.video)

Lipsync Parameters¶

Parameter	Required	Notes
`video_url`	✅	Source video with a speaking face
`audio_url`	—	Audio to sync (mutually exclusive with `text`+`voice_id`)
`text`	—	Text for built-in TTS (requires `voice_id`)
`voice_id`	—	TTS voice: `Emily`, `James`, `Isabella`, `Liam`, `Chloe`, `Adrian`, `Harper`, `Ava`, `Sophia`, `Julia`, `Mason`, `Jack`, `Oliver`, `Ethan`, `Auto`