LTX 2.3 Quality: Reference Video to Video

Try LTX 2.3 Quality: Reference Video to Video in the Workbench

Run this model interactively, tune parameters, and compare outputs.

Model ID: ltx-2-3-quality-reference-video-to-video LTX 2.3 Quality (Reference Video to Video) is the structure-guided preset of Lightricks LTX-2.3 on fal, generating video with synchronized native audio from a reference video, a text prompt, and an optional reference image. It runs the official Union IC-LoRA workflow, where the reference video supplies motion and structure that guide the generated output. The model offers fine control over how closely the result follows the source: video_strength adjusts how much freedom the model has to change the reference motion/structure, while an optional pre-computed control video (depth/edge/pose) can be fed directly into the Union Control conditioning by skipping the built-in control estimation. A true video-to-video mode (preserve_original_video) VAE-encodes the base video as the starting latent, with the denoise strength controlling how much of the original pixels are kept. An optional reference image anchors style or character via the IC-LoRA Union node.

Metric	Value
Parameter Count	22 billion
Mixture of Experts	No
Context Length	Unknown
Multilingual	Unknown
Quantized*	Unknown

*Quantization is specific to the inference provider and the model may be offered with different quantization levels by other providers.

Example request

Use the Workbench as a request builder: configure parameters for this model in the UI, then open the API tab to copy the exact cURL or Python call.

Sync
Async
Async with SSE

This blocks until the video is ready (typically 5-15 minutes). Prefer Async or Async with SSE for anything beyond quick experimentation.See the video generation reference for more details.

Minimal
Basic parameters
All parameters

curl -X POST https://hub.oxen.ai/api/ai/videos/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4"
}'

curl -X POST https://hub.oxen.ai/api/ai/videos/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png"
}'

curl -X POST https://hub.oxen.ai/api/ai/videos/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png",
  "skip_control_preprocess": false,
  "preserve_original_video": false,
  "video_strength": 0.6,
  "strength": 1,
  "num_frames": 121,
  "resolution": "auto",
  "frames_per_second": 24,
  "generate_audio": true,
  "video_quality": "high",
  "negative_prompt": "color distortion, overexposure, static, blurry details, subtitles, style, artwork, painting, frame, still, dim overall tone, worst quality, low quality, JPEG compression artifacts, ugly, mutilated, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, malformed limbs, fused fingers, motionless frame, cluttered background, three legs, crowded background, walking backwards"
}'

See the async queue reference for more details.

Minimal
Basic parameters
All parameters

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4"
}' | jq -r '.generations[0].generation_id')

# Poll until the generation reaches a terminal status.
while true; do
  STATUS=$(curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
    "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq -r '.status')
  echo "Status: $STATUS"
  case $STATUS in succeeded|failed|cancelled) break;; esac
  sleep 5
done

# Print the result.
curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
  "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq .

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png"
}' | jq -r '.generations[0].generation_id')

# Poll until the generation reaches a terminal status.
while true; do
  STATUS=$(curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
    "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq -r '.status')
  echo "Status: $STATUS"
  case $STATUS in succeeded|failed|cancelled) break;; esac
  sleep 5
done

# Print the result.
curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
  "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq .

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png",
  "skip_control_preprocess": false,
  "preserve_original_video": false,
  "video_strength": 0.6,
  "strength": 1,
  "num_frames": 121,
  "resolution": "auto",
  "frames_per_second": 24,
  "generate_audio": true,
  "video_quality": "high",
  "negative_prompt": "color distortion, overexposure, static, blurry details, subtitles, style, artwork, painting, frame, still, dim overall tone, worst quality, low quality, JPEG compression artifacts, ugly, mutilated, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, malformed limbs, fused fingers, motionless frame, cluttered background, three legs, crowded background, walking backwards"
}' | jq -r '.generations[0].generation_id')

# Poll until the generation reaches a terminal status.
while true; do
  STATUS=$(curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
    "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq -r '.status')
  echo "Status: $STATUS"
  case $STATUS in succeeded|failed|cancelled) break;; esac
  sleep 5
done

# Print the result.
curl -s -H "Authorization: Bearer $OXEN_API_KEY" \
  "https://hub.oxen.ai/api/ai/queue/$GEN_ID" | jq .

See the async queue reference for more details.

Minimal
Basic parameters
All parameters

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4"
}' | jq -r '.generations[0].generation_id')

# Stream the SSE channel, grab the data line that follows a
# media_generation_completed event for our id, and pretty-print it.
curl -sN -H "Authorization: Bearer $OXEN_API_KEY" https://hub.oxen.ai/api/events \
  | awk -v id="$GEN_ID" '
    /^event: media_generation_completed$/ { expect=1; next }
    /^data: / && expect {
      payload = substr($0, 7)
      if (index(payload, "\"generation_id\":\"" id "\"")) { print payload; exit }
      expect = 0
    }
  ' | jq .

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png"
}' | jq -r '.generations[0].generation_id')

# Stream the SSE channel, grab the data line that follows a
# media_generation_completed event for our id, and pretty-print it.
curl -sN -H "Authorization: Bearer $OXEN_API_KEY" https://hub.oxen.ai/api/events \
  | awk -v id="$GEN_ID" '
    /^event: media_generation_completed$/ { expect=1; next }
    /^data: / && expect {
      payload = substr($0, 7)
      if (index(payload, "\"generation_id\":\"" id "\"")) { print payload; exit }
      expect = 0
    }
  ' | jq .

# Enqueue, capture the generation id.
GEN_ID=$(curl -s -X POST https://hub.oxen.ai/api/ai/queue \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OXEN_API_KEY" \
  -d '{
  "model": "ltx-2-3-quality-reference-video-to-video",
  "prompt": "<prompt>",
  "input_video": "https://hub.oxen.ai/api/repos/ox/Oxen-AI-Assets/file/main/images/winter_summer_ox.mp4",
  "input_image": "https://hub.oxen.ai/api/repos/elau/assets/file/main/bloxy/bloxy_cropped_512x512.png",
  "skip_control_preprocess": false,
  "preserve_original_video": false,
  "video_strength": 0.6,
  "strength": 1,
  "num_frames": 121,
  "resolution": "auto",
  "frames_per_second": 24,
  "generate_audio": true,
  "video_quality": "high",
  "negative_prompt": "color distortion, overexposure, static, blurry details, subtitles, style, artwork, painting, frame, still, dim overall tone, worst quality, low quality, JPEG compression artifacts, ugly, mutilated, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, malformed limbs, fused fingers, motionless frame, cluttered background, three legs, crowded background, walking backwards"
}' | jq -r '.generations[0].generation_id')

# Stream the SSE channel, grab the data line that follows a
# media_generation_completed event for our id, and pretty-print it.
curl -sN -H "Authorization: Bearer $OXEN_API_KEY" https://hub.oxen.ai/api/events \
  | awk -v id="$GEN_ID" '
    /^event: media_generation_completed$/ { expect=1; next }
    /^data: / && expect {
      payload = substr($0, 7)
      if (index(payload, "\"generation_id\":\"" id "\"")) { print payload; exit }
      expect = 0
    }
  ' | jq .

Fetch model details

The models endpoint returns the full model object, including its json_request_schema.

curl -H "Authorization: Bearer $OXEN_API_KEY" https://hub.oxen.ai/api/ai/models/ltx-2-3-quality-reference-video-to-video

Request parameters

Required parameters

Field	Type	Default	Description
`prompt`	`string`	—	The prompt to guide generation.
`input_video`	`string`	—	The URL of the reference video that supplies motion/structure. Format: uri.

Optional parameters

Field	Type	Default	Description
`input_image`	`string`	—	Optional reference image used by the IC-LoRA Union node for style/character anchoring. Format: uri.
`control_video_url`	`string`	—	Optional pre-computed control video (e.g. an already-rendered depth / edge / pose composite). When provided together with skip_control_preprocess, the built-in control estimation is skipped and this video is fed directly into the Union Control conditioning, resampled to the output resolution and frame count. Format: uri.
`skip_control_preprocess`	`boolean`	`false`	Skip the built-in control estimation (depth / edge / pose) and use control_video_url directly as the control signal. Requires control_video_url; ignored if it is not set.
`preserve_original_video`	`boolean`	`false`	True video-to-video: VAE-encode the base video and use it as the sampler’s starting latent, while control_video_url still drives the IC-LoRA control guide. The amount preserved is controlled by strength. Requires skip_control_preprocess, video and control_video_url.
`video_strength`	`number`	`0.6`	Video conditioning strength. Lower values give the model more freedom to change the reference video motion/structure.
`strength`	`number`	`1`	Sampler denoise strength. With preserve_original_video on, lower values keep more of the original video’s pixels (e.g. 0.5 = keep ~50%), 1.0 fully regenerates. Must be greater than 0.
`num_frames`	`integer`	`121`	The number of frames to generate. Range: 9 – 481.
`resolution`	`string`	`"auto"`	The size of the generated video. ‘auto’ keeps the official Union Control workflow’s low-resolution control output size. The output is generated at up to ~720p (shorter side capped at 704px); larger requested sizes are scaled down, preserving the aspect ratio. One of: auto, square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9.
`frames_per_second`	`number`	`24`	Frames per second of the generated video. Range: 1 – 60.
`generate_audio`	`boolean`	`true`	Whether to include audio in the returned video. When disabled, the final MP4 is returned without an audio track.
`video_quality`	`string`	`"high"`	The quality preset of the generated video. One of: low, medium, high, maximum.
`negative_prompt`	`string`	`"color distortion, overexposure, static, blurry details, subtitles, style, artwork, painting, frame, still, dim overall tone, worst quality, low quality, JPEG compression artifacts, ugly, mutilated, extra fingers, poorly drawn hands, poorly drawn face, deformed, disfigured, malformed limbs, fused fingers, motionless frame, cluttered background, three legs, crowded background, walking backwards"`	The negative prompt to steer generation away from.
`seed`	`integer`	—	Random seed for reproducibility. If None, a random seed is chosen.

Try LTX 2.3 Quality: Reference Video to Video in the Workbench

​Example request

​Fetch model details

​Request parameters

​Required parameters

​Optional parameters

Example request

Fetch model details

Request parameters

Required parameters

Optional parameters