Ready to generate your own videos with simple text prompts? 🎬 Learn how to use @ComfyUI and the @ltx_model LTX-2 optimized model on your RTX AI PC with this educational guide linked below.
Video Models
LTX-2 Text And Image To Video Generator
Lightricks LTX-2 is a production-grade AI creative engine that generates cinematic 1080p videos with synchronized native audio. Available in Pro and Fast modes for text-to-video and image-to-video generation, plus a specialized Retake mode for targeted video edits. Create professional videos with smooth motion, consistent lighting, and lip-sync capabilities perfect for ads, social media, and production workflows.
Public
Generate Audio
LTX-2 Text-to-Video Examples
Explore cinematic videos generated from text prompts using LTX-2's Pro and Fast modes. From epic fantasy to sci-fi horror, see how text descriptions transform into stunning 1080p videos with synchronized native audio.
AI Video
Floating City Arrival
Epic fantasy establishing shot with an ornate airship emerging from clouds, approaching a colossal floating city connected by glowing energy bridges.
Prompt
“Breathtaking wide establishing shot. A magnificent, ornate airship slowly emerges from thick clouds, approaching a colossal, impossible city floating amongst the clouds, connected by glowing energy bridges.”
Live PipelineTake 01 / 12
LTX-2 Image-to-Video Examples
Transform still images into dynamic 1080p videos with synchronized audio. See how LTX-2 brings static scenes to life with smooth motion, from epic fantasy duels to thrilling chase sequences.
Source Feeds01 Inputs
Pipeline
00%
Program · On AirAI · Generated
Output
Transcript · 01
Harry Potter Epic Duel
A cinematic, epic fantasy video, bringing the climactic duel to life. Harry Potter and Voldemort locked in their spell clash amidst the ruins of Hogwarts, with brilliant beams of green and gold light.
Live PipelineTake 01 / 05
LTX-2 Retake Video Editing Examples
Perform targeted retakes on any section of your videos. Replace visuals, audio, or both while preserving timing and continuity. Perfect for fine-tuning specific moments without regenerating the entire video.
Source Feeds01 Inputs
Pipeline
00%
Program · On AirAI · Generated
Output
Transcript · 01
Day to Night Transformation
Change the time in the video from daytime to nighttime.
LTX-2 YouTube Videos
Watch demonstrations and tutorials of LTX-2, the first open-source audio-video generation model
- LTX-2 For Low V-RAM: Audio-Video Model Using Comfy UI (720p & 1080p Videos) - AI Ninja
- LTX-2 Video Models Explained - LTX Studio
- The First Open Source Audio-Video Model — LTX-2 - LTX
- New #1 open-source AI video generator is here! Fast + 4K + audio + low vram - AI Search
- Get Started with LTX-2 in ComfyUI - ComfyUI
- LTX-2 - the first open-weights speech+video generative model - AI Bites
- LTX 2 explained in 8 minutes (Best Open-Source Video Model) - RandomAI
LTX-2 YouTube Videos
Watch demonstrations and tutorials of LTX-2, the first open-source audio-video generation model
LTX-2 Popular Reviews on X
See what the AI community is saying about LTX-2, the first open-source audio-video generation model
RTX AI PCs can now generate videos fast using @ltx_model LTX-2, optimized for @ComfyUI and RTX GPUs. ⚡ Catch up on this and the rest of our CES 2026 news below.
🚀 LTX-2 is now open source: text → audio + video. Today we’re releasing LTX-2, the first open-source foundation model for joint audiovisual generation, together with a full technical report. 🧵👇
LTX-2 is natively supported in ComfyUI on Day 0 🎬🔊 The next chapter in controllability for open-source video generation. - Open-source audio-video foundation model - Generates motion, dialogue, SFX, and music together - Canny, Depth & Pose video-to-video control - Show more
The first truly open-source audio-video model. LTX-2 is a DiT-based foundation model with all core video generation capabilities in one unified model. Designed to run locally on consumer GPUs. - text-to-video - image-to-video - and video-to-video modes 100% open-source.
AI video shouldn’t be locked behind closed systems. We’re releasing LTX-2 as a truly open-source AI video model. Here’s @ZeevFarbman (CEO & Co-Founder, Lightricks) on why openness, local access, and community matter. 🧵
LTX-2 by @Lightricks: Open-source • Native 4K video at up to 50 FPS • High-quality long-form video with strong temporal consistency • Audio-conditioned generation with motion synced to sound • Fine-grained camera & motion control • Efficient enough to run on consumer GPUs Show more
We partnered with @NVIDIA_AI_PC and @Lightricks to push local AI video forward. NVFP4 and NVFP8 checkpoints are now available for LTX-2. With NVIDIA-optimized ComfyUI, LTX-2 delivers cloud-class 4K video locally - up to 3X faster with 60% less VRAM using NVFP4. Show more
LTX-2 is natively supported in ComfyUI on Day 0 🎬🔊 The next chapter in controllability for open-source video generation. - Open-source audio-video foundation model - Generates motion, dialogue, SFX, and music together - Canny, Depth & Pose video-to-video control -
Reel · Specifications
What is LTX-2?
Lightricks' production-grade video AI model with synchronized native audio generation
- · 011080pFull HD Resolution
- · 026-20 secondsVideo Duration Range
- · 033 ModesPro, Fast & Retake
- · 04Native AudioNative AudioSynchronized Sound
First open-source model to achieve cinematic 1080p video with contextually appropriate audio in a unified system.
Reel · Capabilities
LTX-2 Features
Discover the powerful capabilities of Lightricks LTX-2 for professional video production
- Feature 01 / 08
Synchronized Native Audio
Generate cinematic videos with automatically synchronized audio that perfectly matches the visual content. LTX-2 creates immersive soundscapes with dialogue and lip-sync capabilities ideal for ads and social media.
- Feature 02 / 08
Pro & Fast Modes
Choose between Pro mode for highest quality 1080p output or Fast mode for rapid iterations up to 20 seconds. Both modes deliver production-grade results optimized for different workflow needs.
- Feature 03 / 08
Text-to-Video Generation
Transform text prompts into cinematic 1080p videos with smooth, realistic motion and consistent lighting. Create professional video content from descriptions with duration options from 6 to 20 seconds.
- Feature 04 / 08
Image-to-Video Animation
Bring still images to life with natural motion and synchronized audio. LTX-2 maintains original composition and style while adding dynamic movement that feels authentic and engaging.
- Feature 05 / 08
Video Retake Mode
Perform targeted retakes on any video section, replacing visuals, audio, or both while preserving timing and continuity. Perfect for iterative refinement and creative experimentation without full regeneration.
- Feature 06 / 08
Production-Ready Quality
Output 1080p high-definition videos suitable for professional use across advertising, social media, and content production. Consistent quality with realistic motion physics and natural lighting.
- Feature 07 / 08
Flexible Duration Control
Generate videos from 6 to 20 seconds with granular control over timing. Pro mode offers 6-10 second options for premium quality, while Fast mode extends to 20 seconds for longer content.
- Feature 08 / 08
Multi-Modal Capabilities
Three powerful modes in one engine: text-to-video for concept creation, image-to-video for animation, and retake mode for refinement. Comprehensive toolkit for complete video production workflows.
LTX-2 Pro mode delivers the highest quality cinematic videos at 6 credits per second with durations of 6, 8, or 10 seconds. Fast mode prioritizes speed and cost-efficiency at 4 credits per second, supporting flexible durations from 6 to 20 seconds. Both modes generate 1080p videos with synchronized native audio.
Yes, LTX-2 generates synchronized native audio by default for all video outputs. This includes ambient sounds, effects, and audio that matches the visual content. You can disable audio generation if you prefer silent videos or plan to add custom audio later.
Retake mode allows you to perform targeted edits on specific sections of an existing video. You can replace the visuals only, audio only, or both simultaneously while preserving the original timing and continuity. It's perfect for fine-tuning specific moments without regenerating the entire video. Billing is based on actual clip duration, capped at 16 seconds maximum.
Absolutely. LTX-2 supports both text-to-video and image-to-video generation in both Pro and Fast modes. Text-to-video creates videos entirely from your written prompts, while image-to-video animates your still images into dynamic video content with motion and audio.
LTX-2 generates videos in 1080p resolution (Full HD) with cinematic quality. The output format is MP4, which is widely compatible across all platforms and devices. For image-to-video, we accept JPG, PNG, and WEBP input formats up to 10MB.
Pricing is based on video duration: Fast mode costs 4 credits per second, Pro mode costs 6 credits per second, and Retake mode costs 10 credits per second. For Retake, billing is based on actual clip duration with a 16-second maximum cap. Credits are deducted after successful generation.
How to Use LTX-2 Text-to-Video
Choose Your Mode
Select between Pro mode for maximum cinematic quality (6-10 seconds at 6 credits/sec) or Fast mode for quick iterations and cost efficiency (6-20 seconds at 4 credits/sec). Pro mode is ideal for final production assets, while Fast mode is perfect for creative exploration and draft versions.
How to Use LTX-2 Image-to-Video
Upload Your Image
Start by uploading a high-quality still image (JPG, PNG, or WEBP format, up to 10MB). Choose an image with clear composition and good lighting for best results. Select your preferred mode: Pro mode for cinematic quality (6-10 seconds at 6 credits/sec) or Fast mode for quick turnaround (6-20 seconds at 4 credits/sec).
Pricing · Choose Yours
Flexible AI Pricing
Pay-as-you-go credits or subscription plans. No hidden fees, cancel anytime.
One Time supports crypto payment (BTC, USDT, ETH, 350+)
Monthly billing