Kling (via Fal.ai)¶

Kuaishou's Kling v2.6 and o1 models hosted on Fal.ai.

Quick Example¶

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

config = VideoGenerationConfig(
    provider="fal",
    model="fal-ai/kling-video/v2.6",
    api_key="YOUR_FAL_KEY",
)

request = VideoGenerationRequest(
    prompt="A cat playing piano in a jazz club",
    duration_seconds=5,
    aspect_ratio="16:9",
)

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

Supported Models¶

Model prefix	Duration	Image-to-Video	Notes
`fal-ai/kling-video/v2.6`	`5s`, `10s`	✅	Motion control, `cfg_scale`, last-frame pinning
`fal-ai/kling-video/o1`	`5s`, `10s`	✅	Reference-to-video, video edit, start/end frame

Parameters¶

Parameter	Required	Notes
`prompt`	✅
`duration_seconds`	—	`5` or `10`
`image_list` (reference)	—	Image-to-video (kling-o1)
`image_list` (first_frame)	—	First frame pinning (kling-o1)
`image_list` (last_frame)	—	Last frame pinning (kling-v2.6, kling-o1)
`negative_prompt`	—	kling-v2.6 only
`generate_audio`	—	kling-v2.6 only

Motion Control¶

# Kling v2.6 motion control
request = VideoGenerationRequest(
    prompt="A dancer performing on stage",
    duration_seconds=5,
    extra_params={
        "cfg_scale": 0.5,
        "character_orientation": "front",
        "keep_original_sound": True,
    },
)

# Kling o1 reference-to-video elements
request = VideoGenerationRequest(
    prompt="The character walks through a park",
    extra_params={
        "elements": [{"frontal_image_url": "https://example.com/character.jpg"}]
    },
)