APIXO identifies the route specifically as LTX-2 19B from Lightricks. Its public page does not identify the hosted checkpoint as dev, distilled, quantized, or a later LTX-2 variant, so those checkpoint labels should not be applied to this endpoint.
LTX-2 19B by Lightricks is a joint audio-video foundation model available through APIXO for text-to-video, single-image animation, and LoRA-conditioned generation. Create 5–20-second outputs at 480p, 720p, or 1080p, with prompt-based direction for scene composition, motion, camera behavior, and audio cues.
APIXO Workflows
Base / LoRA
Price
From $0.012 / SecBase 1080p is $0.024 / second; LoRA 1080p is $0.03 / second.
Resolution
480P / 720P / 1080P
Duration
5–20 Seconds
Open Weights
Jan 5, 2026
Create with LTX-2 19B
Loading workspace...
A Unified Model for Visual Motion and Sound
Joint Audio-Video Modeling
LTX-2 uses coupled video and audio streams rather than treating sound as a separate post-production stage. This architecture is designed to align visual events with speech, ambience, and Foley described in the prompt.
Temporal Synchronization
Bidirectional cross-modal attention connects the model’s visual and audio timelines. This helps generated sound follow scene events, character activity, environmental changes, and the intended emotional tone throughout a clip.
Detailed Scene Direction
Prompts can describe subjects, actions, composition, lighting, camera movement, and audio cues together. The unified instruction space makes LTX-2 useful for shots where visual staging and sound design must support the same creative idea.
Image-Conditioned Motion
APIXO’s image modes animate exactly one public source image. The input establishes the visual subject and opening composition, while the required prompt directs movement, camera behavior, atmosphere, and scene development.
LoRA-Conditioned Generation
APIXO exposes dedicated text and image modes that accept up to three LoRA items. Each item combines a model path with a scale, enabling specialized visual conditioning without presenting community checkpoints as part of the base model.
Seed-Controlled Iteration
APIXO exposes an optional seed for controlled reruns. Teams can preserve a useful starting condition while refining prompts, resolution, duration, or LoRA strength, although identical seeds do not guarantee identical creative results after other inputs change.
Choose Base Generation or LoRA-Conditioned Production
Base Text and Image Workflows
Generate directly from a prompt or animate one source image without supplying a LoRA. Base modes provide the lowest APIXO rates and suit storyboards, campaign concepts, product motion, and general-purpose audiovisual scene generation.
Text and Image LoRA Workflows
Add up to three LoRA items when a production requires supported specialized conditioning. LoRA modes use the same 5–20-second duration range and resolution tiers but carry separate per-second APIXO rates.
LTX-2 19B Controls on APIXO
Verified generation options exposed by the current public route.
Text / One Image
Source Input
Base / LoRA
Mode Families
5–20 Seconds
Duration
480P / 720P / 1080P
Resolution
16:9 / 9:16
Text Ratios
Up to 3
LoRA Items
Production Workflows for LTX-2 19B
Sound-Aware Storyboards
Develop short sequences in which camera movement, environmental action, dialogue cues, ambience, and Foley belong to the same creative brief. Teams can use the resulting clips to communicate timing and audiovisual intent before committing to a full shoot or animation pipeline.
Product Motion Concepts
Animate an approved product still or generate a scene from text for launch concepts, ecommerce placements, and advertising drafts. Resolution tiers support inexpensive early exploration followed by more detailed outputs for selected directions.
LoRA-Guided Creative Series
Apply supported LoRA conditioning to repeated campaign concepts that require a specialized visual treatment. Seed control helps teams conduct structured, reproducible iterations while retaining direct control over the prompt and source image.
Vertical and Landscape Scenes
Generate text-led 16:9 or 9:16 clips for presentations, social channels, trailers, and editorial concepts. The continuous 5–20-second duration range provides more timing flexibility than endpoints restricted to a few fixed presets.
LTX-2 19B Requirements and Licensing
Important Notes
APIXO requires a prompt of no more than 5,000 characters in every exposed mode.
Image-to-video and image-to-video-LoRA each require exactly one publicly accessible image URL.
Duration must be a whole number from 5 through 20 seconds.
Each LoRA mode accepts up to three items, and every item requires a path and scale.
APIXO does not publicly expose negative prompts, uploaded audio, guidance, frame-rate, first-frame-plus-last-frame, extension, or video-edit controls.
Complex anatomy, embedded text, rapid contact, crowded action, and long continuous narratives may require several generations or separate shots.
Frequently Asked Questions
LTX-2 19B is officially a joint audio-video model, and APIXO asks users to include audio cues in the generation prompt. The public route does not expose an uploaded-audio field, separate soundtrack mode, or control for disabling generated sound.
Base generation costs $0.012, $0.016, or $0.024 per output second at 480p, 720p, or 1080p. Corresponding LoRA-mode rates are $0.015, $0.02, and $0.03 per output second.
Lightricks publishes the LTX-2 weights under the LTX-2 Community License Agreement rather than an unrestricted Apache license. The agreement includes use restrictions and commercial-license requirements for certain organizations that directly use, host, modify, or distribute the model or its derivatives. APIXO-hosted API usage is governed by APIXO’s terms and any applicable provider agreements.
APIXO exposes LoRA paths and scales, but public availability of a file does not establish compatibility, authorization, or commercial rights. Confirm that each LoRA targets the exact supported LTX-2 19B implementation and that its license permits the intended use.
Explore Other Models
Discover more AI models for your next creative workflow