Video

LTX-2.3 (V2V Reference)

Credit-based pricing. Supports image-to-video, reference video, audio generation.

Try LTX-2.3 (V2V Reference) →Compare models

Output

video

Aspect ratio

16:9, 9:16, 1:1

Resolution

1080p, 1440p, 2160p

Duration

3s, 4s, 5s, 6s, 7s, 8s, 10s

Inputs

Reference video (required)

Start image (optional)
Reference audio (optional)
Generates audio

Example prompt

Upload a video to use as motion reference — LTX-2.3 follows the structure and pacing while rendering your scene in any cinematic style

Try this prompt →

LLM-ready

API & LLM schema

Exact request contract for this model. Agents can fetch it from /api/v1/models?id=fal-ai/ltx-2.3-22b/reference-video-to-video.

POST/api/v1/generate48 fields · 3 required

Field	Type	Requirement	Contract
`model_id`	constant	Required	ArtEmotion model identifier.
`extra`	object	Optional	Model-specific settings may also be nested here.
`max_credits`	number	Optional	Reject before submission if the estimated list price exceeds this cap. · Range: 1–…
`webhook_url`	string	Optional	Format: uri
`webhook_secret`	string	Optional	Optional model input.
`folder_id`	string	Optional	Optional model input.
`prompt`	string	Required	Optional model input.
`fps`	number	Optional	The frames per second of the generated video. · Default: 24 · Range: 1–60
`preprocessor`	string	Optional	The preprocessor to use for the generation. · Allowed: depth, canny, pose, none · Default: none
`audio_stg_scale`	number	Optional	The Spatiotemporal Guidance (STG) scale for the audio. Higher values result in more consistent and focused audio content. · Default: 0 · Range: 0–20
`generate_audio`	boolean	Optional	Whether to generate audio for the video. · Default: true
`ic_lora_type`	string	Optional	The IC LoRA type to use for the generation. · Allowed: match_preprocessor, union, detailer, none · Default: union
`use_multiscale`	boolean	Optional	Whether to use multi-scale generation. If True, the model will generate the video at a smaller scale first, then use the smaller video to guide the ge · Default: true
`seed`	integer	Optional	The seed for the random number generator.
`video_strength`	number	Optional	Video conditioning strength. Lower values represent more freedom given to the model to change the video content. · Default: 1 · Range: 0–1
`use_restart_sampling`	boolean	Optional	Whether to use restart sampling. This will inject a small amount of noise during each denoising step, which can help improve the quality of the genera · Default: false
`audio_modality_scale`	number	Optional	The modality scale for the audio. Controls the ratio between video and audio modalities. · Default: 3 · Range: 0–10
`audio_strength`	number	Optional	Audio conditioning strength. Lower values represent more freedom given to the model to change the audio content. · Default: 1 · Range: 0–1
`scheduler`	string	Optional	The scheduler to use. · Allowed: ltx2, linear_quadratic, beta · Default: ltx2
`match_video_length`	boolean	Optional	When enabled, the number of frames will be calculated based on the video duration and FPS. When disabled, use the specified num_frames. · Default: true
`distill_lora_first_pass_scale`	number	Optional	The scale of the distill LoRA to use for the first pass. Set to 0 to disable. · Default: 0.2 · Range: 0–1
`gradient_estimation_gamma`	number	Optional	The gamma of gradient estimation during denoising. Set to 0 to disable. · Default: 2 · Range: 0–10
`distill_lora_second_pass_scale`	number	Optional	The scale of the distill LoRA to use for the second and subsequent passes. · Default: 0.5 · Range: 0–1
`num_frames`	number	Optional	The number of frames to generate. · Default: 121 · Range: 9–481
`acceleration`	string	Optional	The acceleration level to use. · Allowed: none, regular, high, full · Default: regular
`video_output_type`	string	Optional	The output type of the generated video. · Allowed: X264 (.mp4), VP9 (.webm), PRORES4444 (.mov), GIF (.gif) · Default: X264 (.mp4)
`video_rescaling_scale`	number	Optional	The rescaling scale for the video. Controls the ratio between classifier-free guidance and spatiotemporal guidance. · Default: 0.7 · Range: 0–1
`video_write_mode`	string	Optional	The write mode of the generated video. · Allowed: fast, balanced, small · Default: balanced
`video_modality_scale`	number	Optional	The modality scale for the video. Controls the ratio between video and audio modalities. · Default: 3 · Range: 0–10
`audio_rescaling_scale`	number	Optional	The rescaling scale for the audio. Controls the ratio between classifier-free guidance and spatiotemporal guidance. · Default: 0.7 · Range: 0–1
`camera_lora_scale`	number	Optional	The scale of the camera LoRA to use. This allows you to control the camera movement of the generated video more accurately than just prompting the mod · Default: 1 · Range: 0–1
`camera_lora`	string	Optional	The camera LoRA to use. This allows you to control the camera movement of the generated video more accurately than just prompting the model to move th · Allowed: dolly_in, dolly_out, dolly_left, dolly_right, jib_up, jib_down, static, none · Default: none
`video_quality`	string	Optional	The quality of the generated video. · Allowed: low, medium, high, maximum · Default: high
`match_input_fps`	boolean	Optional	When true, match the output FPS to the input video's FPS instead of using the default target FPS. · Default: true
`video_stg_scale`	number	Optional	The Spatiotemporal Guidance (STG) scale for the video. Higher values result in more consistent and focused video content. · Default: 0 · Range: 0–20
`video_cfg_scale`	number	Optional	The Classifier-Free Guidance (CFG) scale for the video. Higher values result in more consistent and focused video content. · Default: 3 · Range: 1–20
`num_inference_steps`	number	Optional	The number of inference steps to use. · Default: 40 · Range: 8–50
`negative_prompt`	string	Optional	The negative prompt to generate the video from. · Default: news broadcast, 3d animation, computer graphics, pc game, console game, video game, cartoon, childish, watermark, logo, text, on screen text, subtitles, titles, signature, slowmo, static
`enable_prompt_expansion`	boolean	Optional	Whether to enable prompt expansion. · Default: true
`enable_safety_checker`	boolean	Optional	Run the safety checker to block unsafe content. · Default: true
`audio_cfg_scale`	number	Optional	The Classifier-Free Guidance (CFG) scale for the audio. Higher values result in more consistent and focused audio content. · Default: 7 · Range: 1–20
`aspect_ratio`	string	Optional	Allowed: 16:9, 9:16, 1:1
`resolution`	string	Optional	Allowed: 1080p, 1440p, 2160p
`duration`	number	Optional	Allowed: 3, 4, 5, 6, 7, 8, 10
`image_url`	string	Optional	Format: uri
`end_image_url`	string	Optional	Format: uri
`video_url`	string	Required	Format: uri
`audio_url`	string	Optional	Format: uri

Minimal request example

{
  "model_id": "fal-ai/ltx-2.3-22b/reference-video-to-video",
  "prompt": "Upload a video to use as motion reference — LTX-2.3 follows the structure and pacing while rendering your scene in any cinematic style",
  "video_url": "https://example.com/video"
}

Raw JSON Schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://www.artemotion.ai/api/v1/models?id=fal-ai%2Fltx-2.3-22b%2Freference-video-to-video",
  "title": "LTX-2.3 (V2V Reference) generation request",
  "description": "Request body accepted by POST /api/v1/generate for fal-ai/ltx-2.3-22b/reference-video-to-video.",
  "type": "object",
  "properties": {
    "model_id": {
      "type": "string",
      "const": "fal-ai/ltx-2.3-22b/reference-video-to-video",
      "description": "ArtEmotion model identifier."
    },
    "extra": {
      "type": "object",
      "additionalProperties": true,
      "description": "Model-specific settings may also be nested here."
    },
    "max_credits": {
      "type": "number",
      "minimum": 1,
      "description": "Reject before submission if the estimated list price exceeds this cap."
    },
    "webhook_url": {
      "type": "string",
      "format": "uri",
      "maxLength": 2048
    },
    "webhook_secret": {
      "type": "string",
      "maxLength": 512
    },
    "folder_id": {
      "type": "string"
    },
    "prompt": {
      "type": "string"
    },
    "fps": {
      "title": "Fps",
      "description": "The frames per second of the generated video.",
      "default": 24,
      "type": "number",
      "minimum": 1,
      "maximum": 60
    },
    "preprocessor": {
      "title": "Preprocessor",
      "description": "The preprocessor to use for the generation.",
      "default": "none",
      "type": "string",
      "enum": [
        "depth",
        "canny",
        "pose",
        "none"
      ]
    },
    "audio_stg_scale": {
      "title": "Audio Stg Scale",
      "description": "The Spatiotemporal Guidance (STG) scale for the audio. Higher values result in more consistent and focused audio content.",
      "default": 0,
      "type": "number",
      "minimum": 0,
      "maximum": 20
    },
    "generate_audio": {
      "title": "Generate Audio",
      "description": "Whether to generate audio for the video.",
      "default": true,
      "type": "boolean"
    },
    "ic_lora_type": {
      "title": "IC-LoRA Type",
      "description": "The IC LoRA type to use for the generation.",
      "default": "union",
      "type": "string",
      "enum": [
        "match_preprocessor",
        "union",
        "detailer",
        "none"
      ]
    },
    "use_multiscale": {
      "title": "Use Multiscale",
      "description": "Whether to use multi-scale generation. If True, the model will generate the video at a smaller scale first, then use the smaller video to guide the ge",
      "default": true,
      "type": "boolean"
    },
    "seed": {
      "title": "Seed",
      "description": "The seed for the random number generator.",
      "type": "integer"
    },
    "video_strength": {
      "title": "Video Strength",
      "description": "Video conditioning strength. Lower values represent more freedom given to the model to change the video content.",
      "default": 1,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "use_restart_sampling": {
      "title": "Use Restart Sampling",
      "description": "Whether to use restart sampling. This will inject a small amount of noise during each denoising step, which can help improve the quality of the genera",
      "default": false,
      "type": "boolean"
    },
    "audio_modality_scale": {
      "title": "Audio Modality Scale",
      "description": "The modality scale for the audio. Controls the ratio between video and audio modalities.",
      "default": 3,
      "type": "number",
      "minimum": 0,
      "maximum": 10
    },
    "audio_strength": {
      "title": "Audio Strength",
      "description": "Audio conditioning strength. Lower values represent more freedom given to the model to change the audio content.",
      "default": 1,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "scheduler": {
      "title": "Scheduler",
      "description": "The scheduler to use.",
      "default": "ltx2",
      "type": "string",
      "enum": [
        "ltx2",
        "linear_quadratic",
        "beta"
      ]
    },
    "match_video_length": {
      "title": "Match Video Length",
      "description": "When enabled, the number of frames will be calculated based on the video duration and FPS. When disabled, use the specified num_frames.",
      "default": true,
      "type": "boolean"
    },
    "distill_lora_first_pass_scale": {
      "title": "Distill Lora First Pass Scale",
      "description": "The scale of the distill LoRA to use for the first pass. Set to 0 to disable.",
      "default": 0.2,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "gradient_estimation_gamma": {
      "title": "Gradient Estimation Gamma",
      "description": "The gamma of gradient estimation during denoising. Set to 0 to disable.",
      "default": 2,
      "type": "number",
      "minimum": 0,
      "maximum": 10
    },
    "distill_lora_second_pass_scale": {
      "title": "Distill Lora Second Pass Scale",
      "description": "The scale of the distill LoRA to use for the second and subsequent passes.",
      "default": 0.5,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "num_frames": {
      "title": "Num Frames",
      "description": "The number of frames to generate.",
      "default": 121,
      "type": "number",
      "minimum": 9,
      "maximum": 481,
      "multipleOf": 1
    },
    "acceleration": {
      "title": "Acceleration",
      "description": "The acceleration level to use.",
      "default": "regular",
      "type": "string",
      "enum": [
        "none",
        "regular",
        "high",
        "full"
      ]
    },
    "video_output_type": {
      "title": "Video Output Type",
      "description": "The output type of the generated video.",
      "default": "X264 (.mp4)",
      "type": "string",
      "enum": [
        "X264 (.mp4)",
        "VP9 (.webm)",
        "PRORES4444 (.mov)",
        "GIF (.gif)"
      ]
    },
    "video_rescaling_scale": {
      "title": "Video Rescaling Scale",
      "description": "The rescaling scale for the video. Controls the ratio between classifier-free guidance and spatiotemporal guidance.",
      "default": 0.7,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "video_write_mode": {
      "title": "Video Write Mode",
      "description": "The write mode of the generated video.",
      "default": "balanced",
      "type": "string",
      "enum": [
        "fast",
        "balanced",
        "small"
      ]
    },
    "video_modality_scale": {
      "title": "Video Modality Scale",
      "description": "The modality scale for the video. Controls the ratio between video and audio modalities.",
      "default": 3,
      "type": "number",
      "minimum": 0,
      "maximum": 10
    },
    "audio_rescaling_scale": {
      "title": "Audio Rescaling Scale",
      "description": "The rescaling scale for the audio. Controls the ratio between classifier-free guidance and spatiotemporal guidance.",
      "default": 0.7,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "camera_lora_scale": {
      "title": "Camera LoRA Scale",
      "description": "The scale of the camera LoRA to use. This allows you to control the camera movement of the generated video more accurately than just prompting the mod",
      "default": 1,
      "type": "number",
      "minimum": 0,
      "maximum": 1
    },
    "camera_lora": {
      "title": "Camera LoRA",
      "description": "The camera LoRA to use. This allows you to control the camera movement of the generated video more accurately than just prompting the model to move th",
      "default": "none",
      "type": "string",
      "enum": [
        "dolly_in",
        "dolly_out",
        "dolly_left",
        "dolly_right",
        "jib_up",
        "jib_down",
        "static",
        "none"
      ]
    },
    "video_quality": {
      "title": "Video Quality",
      "description": "The quality of the generated video.",
      "default": "high",
      "type": "string",
      "enum": [
        "low",
        "medium",
        "high",
        "maximum"
      ]
    },
    "match_input_fps": {
      "title": "Match Input Fps",
      "description": "When true, match the output FPS to the input video's FPS instead of using the default target FPS.",
      "default": true,
      "type": "boolean"
    },
    "video_stg_scale": {
      "title": "Video Stg Scale",
      "description": "The Spatiotemporal Guidance (STG) scale for the video. Higher values result in more consistent and focused video content.",
      "default": 0,
      "type": "number",
      "minimum": 0,
      "maximum": 20
    },
    "video_cfg_scale": {
      "title": "Video Cfg Scale",
      "description": "The Classifier-Free Guidance (CFG) scale for the video. Higher values result in more consistent and focused video content.",
      "default": 3,
      "type": "number",
      "minimum": 1,
      "maximum": 20
    },
    "num_inference_steps": {
      "title": "Num Inference Steps",
      "description": "The number of inference steps to use.",
      "default": 40,
      "type": "number",
      "minimum": 8,
      "maximum": 50,
      "multipleOf": 1
    },
    "negative_prompt": {
      "title": "Negative Prompt",
      "description": "The negative prompt to generate the video from.",
      "default": "news broadcast, 3d animation, computer graphics, pc game, console game, video game, cartoon, childish, watermark, logo, text, on screen text, subtitles, titles, signature, slowmo, static",
      "type": "string"
    },
    "enable_prompt_expansion": {
      "title": "Enable Prompt Expansion",
      "description": "Whether to enable prompt expansion.",
      "default": true,
      "type": "boolean"
    },
    "enable_safety_checker": {
      "title": "Enable Safety Checker",
      "description": "Run the safety checker to block unsafe content.",
      "default": true,
      "type": "boolean"
    },
    "audio_cfg_scale": {
      "title": "Audio Cfg Scale",
      "description": "The Classifier-Free Guidance (CFG) scale for the audio. Higher values result in more consistent and focused audio content.",
      "default": 7,
      "type": "number",
      "minimum": 1,
      "maximum": 20
    },
    "aspect_ratio": {
      "type": "string",
      "enum": [
        "16:9",
        "9:16",
        "1:1"
      ]
    },
    "resolution": {
      "type": "string",
      "enum": [
        "1080p",
        "1440p",
        "2160p"
      ]
    },
    "duration": {
      "type": "number",
      "enum": [
        3,
        4,
        5,
        6,
        7,
        8,
        10
      ]
    },
    "image_url": {
      "type": "string",
      "format": "uri"
    },
    "end_image_url": {
      "type": "string",
      "format": "uri"
    },
    "video_url": {
      "type": "string",
      "format": "uri"
    },
    "audio_url": {
      "type": "string",
      "format": "uri"
    }
  },
  "required": [
    "model_id",
    "prompt",
    "video_url"
  ],
  "additionalProperties": false
}

Browse schemas for every ArtEmotion model →

FAQ

How much does LTX-2.3 (V2V Reference) cost on ArtEmotion?

Credit-based pricing. You pay in ArtEmotion credits — every plan and top-up converts USD to credits at a fixed rate.

Do I get my credits back if LTX-2.3 (V2V Reference) fails?

Yes — failed generations are never charged. The credits are released back to your balance automatically.

Can I call LTX-2.3 (V2V Reference) from the API?

Yes. Use POST /api/v1/generate with model_id: "fal-ai/ltx-2.3-22b/reference-video-to-video". See the API reference for the full schema.

Where are my generations stored?

Every output is saved to your personal Library. You can export or delete everything any time from Privacy & deletion.

Ready to generate with LTX-2.3 (V2V Reference)?

Start now →

No commitment. New accounts get 50 free credits on signup. See pricing.