# LTX

> LTX builds open world models: AI that understands and generates physical reality across video, audio, and multimodal inputs. The LTX ecosystem includes open-weight models (18M+ downloads on HuggingFace), a managed API, LTX Studio, and LTX Desktop. Built for AI teams, developers, and production pipelines. Open weights, full local deployment, no vendor lock-in, and roughly one-eighth the cost of comparable closed models.

LTX focuses on controllable, production-ready AI video and audio generation.
The ecosystem includes open-source model weights, a developer API for programmatic video generation, LTX Desktop for local deployment, and LTX Studio for creative production teams.

LTX is designed for AI-native platforms, enterprise teams, animation studios, VFX houses, and developers who need long-form, high-resolution, production-ready outputs that integrate into real-world pipelines.

## Agentic Access

LTX does not currently expose a dedicated MCP server. AI agents that need to take action — generating video/audio, checking model versions, or fine-tuning a custom model — should use the following endpoints directly rather than crawling the marketing pages below:

- [API Documentation](https://docs.ltx.io/welcome): Canonical, actively maintained reference for calling the LTX API programmatically — endpoints, request/response formats, and auth. Treat this as the source of truth for API behavior, since it updates independently of the pages below. The docs site publishes its own `/llms.txt` index at the root, and appends `.md` to any page URL for a clean markdown version — agents should use those directly rather than parsing rendered HTML.
- [API Overview](https://ltx.io/model/api): High-level entry point to the API for agents that need context before reading the full docs.
- [API Pricing](https://ltx.io/model/api/pricing): Usage-based pricing, billed per second of video generated — check before executing any paid generation task on a user's behalf.
- [Developer Console](https://console.ltx.video/): Where agents (or the humans directing them) sign up for an API key and can test calls in the Playground before integrating.
- [GitHub](https://github.com/Lightricks/LTX-2): Source code, open model weights, issue tracker, and the official ComfyUI integration for direct/local integration.
- [HuggingFace](https://huggingface.co/Lightricks/LTX-2.3): Model weights, versions, and inference-ready checkpoints agents can pull directly.
- [LTX Trainer Documentation](https://docs.ltx.io/open-source-model/ltx-trainer/ltx-training): Setup and configuration reference for agents tasked with training a LoRA or IC-LoRA on custom characters, styles, or motion.
- [Model License](https://ltx.io/model/license): Terms agents should check before recommending or executing commercial use — free under $10M ARR, commercial license required above that threshold.

## Core Pages

- [Homepage](https://ltx.io/): Overview of the LTX ecosystem, models, and developer offerings.
- [Models](https://ltx.io/model): Overview of available LTX models and their capabilities.
- [About LTX](https://ltx.io/about-us): Background on LTX and the team behind the models.

## Models

- [LTX-2](https://ltx.io/model/ltx-2): Production-grade model page for the LTX-2 family.
- [LTX-2.3](https://ltx.io/model/ltx-2-3): Latest LTX release with native portrait video, audio-to-video, and Fast and Pro tiers.
- [LTXV](https://ltx.io/model/ltxv): Overview of the LTXV model variant.
- [Video Generation Model](https://ltx.io/model/video-generation-model): Technical overview of the LTX video generation model family.
- [LTX Model Alternatives](https://ltx.io/model/alternatives): Comparison pages for LTX-2.3 against competing video generation models.
- [Developer Program](https://ltx.io/model/ltx-developer-program): Program for developers building products and research on top of LTX models.
- [Model License](https://ltx.io/model/license): License terms governing use of LTX models. Free for organizations under $10M ARR; commercial license required above that threshold.

## Capabilities

- [Text to Video](https://ltx.io/model/capabilities/text-to-video): Generate video directly from text prompts with LTX models.
- [Audio to Video](https://ltx.io/model/capabilities/audio-to-video): Generate video driven by audio input with LTX models.
- [Image to Video](https://ltx.io/model/capabilities/image-to-video-model): Generate high-quality video from a single image or multiple keyframes.
- [Video to Video](https://ltx.io/model/capabilities/video-to-video-model): Transform existing footage with precise control over structure, motion, and continuity.

## API

- [API Overview](https://ltx.io/model/api): Documentation and guidance for integrating LTX models via API.
- [API Pricing](https://ltx.io/model/api/pricing): Pricing details and usage tiers for the LTX API. Usage-based, billed per second of video generated.

## LoRA Training

- [LTX Trainer](https://ltx.io/model/ltx-trainer): Official open-source training framework for LTX. Train LoRAs and IC-LoRAs on custom characters, styles, and motion. Supports video, audio, and cross-modal conditioning modes in a single framework.
- [LTX Trainer Documentation](https://docs.ltx.io/open-source-model/ltx-trainer/ltx-training): Full documentation for LTX Trainer setup, configuration, and training runs.
- [Training Your First LoRA on LTX](https://ltx.io/blog/training-your-first-lora-on-ltx): Step-by-step guide to running a first LoRA training job with the LTX Trainer.

## Research and Open Source

- [LTX-2.3 Open Source](https://ltx.io/model/open-source): Details on the open-source release and model weights of LTX-2.3.
- [LTX Community Library](https://ltx.io/ltx-community): Curated library of community-built LoRAs, IC-LoRAs, ComfyUI workflows, and products built on LTX. Includes submission flow for creators.
- [Academic Grants](https://ltx.io/model/grants): Information on research grants and academic collaboration programs.
- [LTX-2 Technical Report (PDF)](https://cdn.prod.website-files.com/68872d15af29880764eac4aa/695c06aa63b560e217a68363_LTX_2_Technical_Report_compressed.pdf): Full technical report on LTX-2 architecture, training, and performance.
- [LTX-2 Research Paper (arXiv)](https://arxiv.org/abs/2601.03233): Peer-reviewed research paper for LTX-2.

## Blog

- [Blog](https://ltx.io/blog): Main LTX blog covering product updates, tutorials, prompting guides, and production practices.
- [Introducing the New LTX Trainer](https://ltx.io/blog/introducing-the-new-ltx-trainer-one-framework-every-training-mode): Announcement of the new LTX Trainer supporting video, audio, and cross-modal conditioning modes (June 2026).
- [LTX-2.3 Release](https://ltx.io/model/model-blog/ltx-2-3-release): Release notes and overview of LTX-2.3.
- [LTX-2.3 Prompt Guide](https://ltx.io/model/model-blog/ltx-2-3-prompt-guide): Prompting guidance specific to LTX-2.3.
- [LTX-2.3 Fast vs Pro](https://ltx.io/model/model-blog/ltx-2-3-fast-vs-pro): Comparison of the Fast and Pro tiers of LTX-2.3.
- [LTX-2.3 Portrait Video](https://ltx.io/model/model-blog/ltx-2-3-portrait-video): Guide to generating native portrait video with LTX-2.3.
- [LTX-2 Image to Video and Text to Video Workflow](https://ltx.io/model/model-blog/ltx-2-image-to-video-text-to-video-workflow): Walkthrough of LTX-2 workflows across input modes.
- [How to Generate 20 Second AI Videos](https://ltx.io/model/model-blog/how-to-generate-20-second-ai-videos): Guide to producing longer-form AI video generations with LTX.
- [How to Reduce Warble and AI Pattern Artifacts](https://ltx.io/model/model-blog/how-to-reduce-warble-and-ai-pattern-artifacts-in-ltx-2): Techniques for cleaning up common artifacts in LTX-2 outputs.
- [How to Set Up LTX Desktop](https://ltx.io/model/model-blog/how-to-set-up-ltx-desktop): Setup guide for the LTX Desktop application.
- [ComfyUI Workflow Guide](https://ltx.io/model/model-blog/comfyui-workflow-guide): Using LTX models inside ComfyUI workflows.
- [AI Video Generation at Scale](https://ltx.io/model/model-blog/ai-video-generation-at-scale): Practices for running LTX video generation at production scale.

## Newsroom

- [Newsroom](https://ltx.io/newsroom): All LTX press releases, announcements, and media resources.
- [Introducing the New LTX Trainer](https://ltx.io/newsroom/introducing-the-new-ltx-trainer-one-framework-every-training-mode): Press release on the LTX Trainer launch (June 17, 2026).
- [LTX-2 Foundation Model Released](https://ltx.io/newsroom/ltx-2-foundation-ai-video-model-is-released): Press release on the launch of LTX-2 as a foundation video model.
- [LTX-2 is Now Open Source](https://ltx.io/newsroom/ltx-2-is-now-open-source-full-model-weights-released): Press release on the full LTX-2 model weights being open sourced.
- [Artificial Analysis Ranks LTX-2 Among Top AI Video Models](https://ltx.io/newsroom/artificial-analysis-ranks-ltx-2-among-top-ai-video-models): Third-party benchmarking coverage ranking LTX-2 among the top AI video models.
- [LTX and the Democratization of Creative Production](https://ltx.io/newsroom/ltx-and-the-democratization-of-creative-production): Feature on LTX's role in democratizing creative production.

## Glossary

- [LTX Glossary](https://ltx.io/glossary): Reference definitions for key AI video generation concepts used across LTX, including diffusion models, LoRA, ControlNet, VAE, flow matching, and more.

## Documentation

- [LTX Documentation](https://docs.ltx.io/welcome): Comprehensive documentation covering models, APIs, and usage.

## Platform

- [LTX Studio](https://ltx.io/studio): Creative platform for AI video production. Built for agencies, brand studios, and film and TV teams. Uses LTX-2.3 alongside other models including Kling, Veo, FLUX, and Seedance.
- [LTX Desktop](https://ltx.io/ltx-desktop): Free, open-source desktop application for running LTX-2.3 fully locally. No cloud dependency, no per-generation fees, no data leaving the machine.

## Community

- [GitHub](https://github.com/Lightricks/LTX-2): Source code, issues, and contributions for LTX-2 and LTX-2.3.
- [HuggingFace](https://huggingface.co/Lightricks/LTX-2.3): Model weights, community discussions, and version history.
- [LTX Community Library](https://ltx.io/ltx-community): Curated LoRAs, IC-LoRAs, ComfyUI workflows, and products built by the community.
- [ComfyUI Nodes](https://github.com/Lightricks/ComfyUI-LTXVideo): Official ComfyUI integration for LTX models.
- [Discord](https://discord.com/invite/ltxplatform): Official LTX community server for developers, researchers, and creators.
- [X / Twitter](https://x.com/ltx_model): Official LTX account for model releases, updates, and announcements.

## Support and Contact

- [Help Center](https://help.ltx.io/hc/en-us): Support articles and troubleshooting for LTX products.
- [Trust Center](https://trust.lightricks.com): Security, privacy, and compliance information.
- [Privacy Policy](https://static.lightricks.com/legal/Privacy%20Policy%20-%20LTX%20Platform.pdf): LTX Platform privacy policy.
- [Contact Sales](https://ltx.io/forms/ltx-contact-sales): Enterprise licensing and commercial sales inquiries.