Malik MuzamilMalikMuzamil.ai Join our community
You commented LTX

Run LTX-2.5 AI video on your own computer, or the cheapest route if your machine can't

This is the step-by-step guide from the reel: the official download links, the exact hardware each route needs (PC and Mac), three ways to install it, the settings and prompt for your first clip with sound, and what it costs when you rent a GPU instead. Everything was checked on the official LTX, GitHub and Hugging Face pages on October 28, 2026.

CostFree model · cloud from $0.50/hr
Time15 min first clip · download 1 to 8 hrs
LevelBeginner (Route A) to coder (Route C)
Works onNVIDIA PC, Apple Silicon Mac, cloud
TONIGHT1
  • You know which route fits your machine
  • Your first LTX-2.5 clip with sound
  • A prompt formula plus 10 ready prompts
  • The licence answer for client work

IN ONE SENTENCELTX-2.5 is a free, open-weights video model from Lightricks that makes video and matching audio together, and it runs on your own machine if you have an NVIDIA card with about 24 GB of video memory or an Apple Silicon Mac with 32 GB or more.

Before you download 70 GB

Four things the reel simplified

The honest version, checked today

It is LTX-2.5, by Lightricks. The auto captions heard "RTX 2.5". It came out in August 2026 (licence dated August 11) and it is still the newest LTX model on October 28.

Free, with a hardware bill. The model and the apps cost nothing. But a 22 billion parameter video model needs a strong graphics card. No strong card means renting one (about $0.50 to $1.19 an hour) or paying per clip on the LTX API.

"Open source" means open weights with a revenue rule. Anyone, including freelancers doing client work, can use it free while their company makes under $10 million a year. Above that you need a paid licence. It is not a classic open source licence like MIT.

Better prompt following, not perfect. The Gemma 4 text encoder does read prompts better. Each run still uses a new random seed, and LTX's own docs say on-screen text is not guaranteed. Plan on two or three tries.

Visual 1 · Route chooser

Which route fits your machine

Tap what you have. The chart shows how far your machine gets and which route to follow below.

How much video memory (VRAM) does your card have?Windows: Task Manager, Performance, GPU, "Dedicated GPU memory"
24 GB or more
RTX 3090, 4090, 5090
Route A or BFull quality, 1080p, up to 10 s and longer at lower sizes.
12 to 16 GB
RTX 3060 12 GB, 4060 Ti 16 GB
Route A, limited720p clips up to 8 to 10 s. Needs 40 GB RAM (32 GB if models sit on an NVMe drive).
Under 12 GB
Route DRent a GPU or use the free credits. Your card is below the comfortable floor.
Are you on macOS 26 (Tahoe), and how much memory?Apple menu, About This Mac, "Memory"
48 GB or more
macOS 26+
Route A (LTX Desktop)Full local grid except 1080p at 20 s.
32 GB
macOS 26+
Route A, limited540p and 5 seconds only. Good for testing prompts.
Under 32 GB, or macOS 14 / 15
Route DThe app opens, but local generation is off. ComfyUI and the Python repo need NVIDIA.
What do you want to spend?No NVIDIA card or Apple Silicon means no local route
Nothing, just try it
LTX Studio free800 one-time credits, LTX-2.5 included, personal use only.
A few dollars, full control
Route D: RunpodRTX 4090 at $0.89/hr runs the same ComfyUI template as a PC.
Pay per clip, no setup
LTX API$0.09 per second at 720p, so $0.54 for a 6 s clip.
Runs wellRuns with limitsPay to run
Visual 2 · Hardware

How much video memory each size of clip needs

Bars are to scale (80 GB is the full bar). Numbers come from LTX Desktop's own compatibility table and LTX's system requirements page.

NVIDIA · VRAM on the card

Dashed line = ComfyUI + Python official minimum (32 GB)
9 GBfloor
The app will start on an fp8 card (RTX 40 or 50 series). LTX does not recommend it.
12 GBRTX 3060
540p up to 10 s, 720p up to 8 s. No 1080p.
16 GB4060 Ti 16G
540p up to 20 s, 720p up to 10 s, 1080p up to 6 s. Needs 40 GB system RAM.
24 GB3090 / 4090
The full LTX Desktop grid, including 1080p at 10 s. LTX Desktop's stated minimum.
32 GBRTX 5090
LTX Desktop recommended. Official minimum for ComfyUI and the Python repo. LoRA training with the low-VRAM config.
80 GBA100 / H100
Recommended for the full Python pipelines and standard LoRA training. Rent it, do not buy it.

Mac · unified memory, LTX Desktop only, macOS 26+

<32 GBany Mac
No local generation. The app runs in API mode only.
32 GBM-series
540p and 5 seconds.
48 GB+M-series
Full local grid except 1080p at 20 s.
Every route also needs about 100 GB of free disk (the LTX-2.5 core files are about 72 GB) and 32 GB of system RAM.
Client work and the licence

The licence in plain words

LTX-2.5 uses the LTX-2.x Community License Agreement (dated August 11, 2026). The parts that matter for you:

✓
Under $10M a yearFree for any use, including paid client work and ads
✓
Your own LoRAsAllowed. Fine-tunes count as "Derivatives" under the same licence
✕
$10M a year or morePaid Commercial Use Agreement, email licensing@ltx.io
✕
LTX Studio free planDifferent product: personal use only, no business work

Revenue counts the whole company group, not just you. The licence also has a list of banned uses (Attachment A). This is a summary, not legal advice: read the full licence before big client jobs.

What each route really costs

RouteBest forCostGood to know
A · LTX DesktopBeginners with a strong PC or MacFreeWindows, Linux, Apple Silicon. Beta app. Text encoding through the free LTX API key saves memory
B · ComfyUI templateControl, LoRAs, image-to-videoFreeNVIDIA only. Official minimum 32 GB VRAM, the template ships lower-VRAM files
C · LTX-2 repoCoders, batch jobsFreeAbout 66 GB download. Python 3.12+, CUDA 12.7+
D · Runpod GPUNo strong card$0.50 to $1.19/hrRTX 3090 $0.50, RTX 4090 $0.89, RTX 5090 $1.19 per hour. Start with $10 of credit
LTX StudioTesting prompts tonightFree, then $15/mo800 one-time credits. Free plan is personal use only
LTX APIPay per clip$0.09/s at 720pFast model: $0.54 for 6 s at 720p, $0.78 at 1080p. Prepaid credits

Prices checked October 28, 2026. Pakistan and India: Runpod, LTX Studio and the LTX API bill in US dollars, so use a card that allows international online payments. Big downloads: 72 GB takes about 1.5 hours on 100 Mbps and about 8 hours on 20 Mbps.

Your quick win tonight · 30 minutes

Your first LTX-2.5 clip with sound, before the big download finishes

The local download takes hours on most connections. So start it, then test your prompt in the cloud while it runs:

  1. Start the download for your route (Route A step 3, or Route B step 3). Leave it running.
  2. Open LTX Studio at ltx.io/studio and sign up for the Free plan ($0, 800 one-time credits).
  3. Copy the first-clip prompt below, pick the LTX-2.5 model, 16:9, and generate.
  4. Change one thing (the camera move or the spoken line) and generate again. You now know how LTX reads a prompt.
Result: a 5 second clip of a chai seller with his voice and market sound, plus a prompt you trust for the moment your local setup is ready.

Stuck on a step? Post a screenshot in The 1% AI Circle and we'll fix it with you.

first-clip prompt · paste into ltx
A medium close-up shot of a young chai seller in a busy Lahore market at golden hour. He pours steaming tea from a height into a small glass, the stream arcing smoothly, then looks up at the camera and smiles. He wears a faded blue kurta and a white cap, a thin mustache, sleeves rolled up. Warm orange sunlight cuts through dust in the air, string lights blur in the background. The camera slowly pushes in toward his hands as the tea settles. We hear the clink of glass, market chatter and a distant rickshaw horn, and he says in a warm voice, "Garam chai, fresh, fresh!"
It worked if the man pours the tea, looks up, and you hear the glass, the market and his line "Garam chai, fresh, fresh!" in one clip.
Route A · easiest

LTX Desktop: the official app, no code

Lightricks' own free app. It downloads the model, picks settings that fit your hardware and greys out the ones that don't. Status: beta.

Once5 minutes

The LTX-2.5 files are "gated": free, but you must accept the terms first.

  1. Make a free account at huggingface.co.
  2. Open huggingface.co/Lightricks/LTX-2.5 and click the button to agree and get access.
It worked if the model page shows the file list instead of the access form.
Once5 minutes
  1. Go to the latest LTX Desktop release.
  2. Windows: download LTX-Desktop-Setup.exe. Mac (Apple Silicon): LTX-Desktop-arm64.dmg. Linux: the .AppImage or .deb.
  3. Install and open it.
If you see "Coming soon to your device", your machine is under the local floor (less than 9 to 15 GB VRAM, less than 32 GB RAM on a Mac, or macOS older than 26). Use Route D.
Once1 to 8 hours of download
  1. On the Setup screen, sign in to Hugging Face.
  2. Text encoding: paste a free key from the LTX Console (LTX says API text encoding is free and saves memory), or pick the local encoder to stay fully offline (extra download).
  3. Download LTX 2.5 Fast, the default local model. Keep about 100 GB free.
It worked if Settings, Models shows LTX 2.5 Fast as installed.
5 minutes
  1. Open Text to Video from the home gallery.
  2. Paste the first-clip prompt, keep the suggested size and 5 seconds, click generate.
  3. Find it in the Assets library.
It worked if you get an MP4 with sound in Assets.
If you see sizes greyed out with a reason, that size does not fit your memory. Drop to 720p or a shorter clip.
Route B · most control

ComfyUI with the official LTX-2.5 template

ComfyUI is a free node editor. LTX calls it the recommended way for most users. NVIDIA cards only for LTX.

Once10 minutes
  1. Download ComfyUI Desktop from comfy.org/download and install it.
  2. Already have ComfyUI? Update it first. The LTX-2.5 templates need a recent version.
Once2 minutes

Same as Route A step 1: accept access on the LTX-2.5 Hugging Face page. The template itself reminds you to do this first.

Once1 to 8 hours of download
  1. Click Templates, search LTX-2.5.
  2. Pick Text-to-Video (video_ltx2_5_t2v). There is also Image-to-Video and First-Frame / Last-Frame.
  3. Press Download all. It fetches the lower-VRAM (int8) model, the Gemma 4 text encoder, the video and audio decoders and the 2x upscaler.
It worked if the Missing Models list is empty and no node is red.
If you see red nodes, click Install Missing Custom Nodes, then fully restart ComfyUI (close and reopen, not just refresh).
5 minutes

The template defaults are a good first run. Paste your prompt into the prompt field and click Run. Output lands in ComfyUI/output/.

SettingStart withWhy
width x height1280 x 720Template default. Must divide by 32
duration5 sShort clips fit more cards
frame_rate24Cinematic. 25 standard, 30 smoother
seedfixedLets you repeat a result you like
prompt_enhanceoffTurn on for short rough prompts (extra model download, adds 1 to 2 min)
If you see "out of memory": lower the size (try 960 x 544), shorten the clip, close other apps, or move to Route D. Frame counts must be 1 plus a multiple of 8 (97, 121) if you type frames instead of seconds.
Route C · for coders

The official LTX-2 repo from the terminal

Best for batch jobs on a Linux box or a rented GPU. Needs an NVIDIA card (32 GB+ VRAM per LTX), Python 3.12+ and CUDA 12.7+. Commands are from the official README.

Install the two tools first

  1. uv (Python manager): Mac or Linux line below, or the Windows PowerShell line.
  2. hf (Hugging Face command line): same idea.
mac / linux terminal
$curl -LsSf https://astral.sh/uv/install.sh | sh
$curl -LsSf https://hf.co/cli/install.sh | bash
windows powershell
>powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
>powershell -ExecutionPolicy ByPass -c "irm https://hf.co/cli/install.ps1 | iex"
terminal · get the code
$git clone https://github.com/Lightricks/LTX-2.git
$cd LTX-2
$uv sync --extra natten
$hf auth login
The natten extra speeds up decoding on Linux; on Windows and Mac it is skipped by itself. Log in with a Hugging Face Read token.
terminal · download about 66 GB
hf download Lightricks/LTX-2.5 diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors vae/ltx-2.5-video-vae-bf16.safetensors vae/ltx-2.5-audio-vae-bf16.safetensors latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors --local-dir models/ltx-2.5
terminal · generate (mac / linux line breaks)
uv run python -m ltx_pipelines.distilled \
  --transformer-path models/ltx-2.5/diffusion_models/ltx-2.5-22b-distilled-transformer-bf16.safetensors \
  --text-encoder-path models/ltx-2.5/text_encoders/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
  --video-vae-path models/ltx-2.5/vae/ltx-2.5-video-vae-bf16.safetensors \
  --audio-vae-path models/ltx-2.5/vae/ltx-2.5-audio-vae-bf16.safetensors \
  --spatial-upsampler-path models/ltx-2.5/latent_upscale_models/ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors \
  --num-frames 121 --seed 42 --output-path output.mp4 \
  --quantization fp8-cast --offload cpu \
  --prompt "PASTE YOUR PROMPT HERE"
It worked if output.mp4 appears in the LTX-2 folder with sound. 121 frames at 24 fps is about 5 seconds.
If you see a 401 or 403 error on download, accept the terms on the model page and log in with a Read token (fine-grained tokens need "read gated repos" turned on). The last two flags (--quantization fp8-cast --offload cpu) are LTX's own memory savers; remove them on an 80 GB card for speed. On Windows, put the command on one line without the backslashes.
Route D · no strong GPU

Rent a GPU by the hour on Runpod

You get a real NVIDIA card in the browser and run the exact Route B template on it. You pay only while the machine is on.

Once5 minutesPaid

Sign up at runpod.io and add at least $10 of credit (what Runpod's own ComfyUI tutorial asks for).

10 to 30 minutes first start
  1. RTX 4090 or 3090: deploy Runpod's ComfyUI template. RTX 5090: use ComfyUI Blackwell Edition.
  2. Set the volume disk to about 150 GB so the 72 GB of LTX files fit.
  3. Choose On-Demand, click Deploy On-Demand. The first start can take up to 30 minutes.
  4. Click Connect, then Connect to HTTP Service [Port 8188]. ComfyUI opens in your browser.
Every time

Do Route B steps 3 and 4 inside that browser tab. Download your MP4s, then stop the pod in the Runpod console. A running pod keeps billing even when you are asleep.

If you see "Bad Gateway" or "Not Ready" on port 8188, wait 2 to 3 minutes and refresh.

Visual 3 · Cost per clip calculator · time your first clip, type it in

Cost per clip on Runpod$0.04
Clips for $10 of credit224
Same 6 s clip on the LTX API (fast, 720p)$0.54

Math: hourly price x minutes / 60. The 3 minute starting value is an example, not a benchmark: time your own first clip, because speed depends on the card, the size and the length. Setup time, model download and idle time also bill, so real cost per clip is higher on day one.

Other no-GPU options. LTX Studio: free 800 one-time credits (personal use), then Lite $15 a month. LTX API: fast model $0.09 a second at 720p and $0.13 at 1080p, pro model $0.12 and $0.17, clips of 6 to 20 seconds.

Visual 4 · How it works

What happens when you click Run

This is why it needs so much memory, and why the sound always matches the picture: video and audio are made together in one pass, then sharpened.

01Your promptA paragraph: shot, action, camera, sound
02Gemma 4 reads it12B text encoder turns words into instructions
03Stage 1Video and audio made together at half size
042x upscalerDoubles width and height
05Stage 2Refines detail, keeps the audio
06MP4 with soundSaved to your output folder

In the ComfyUI template, a 1280 x 720 clip is drafted at 640 x 360 in stage 1. The 22B model plus the 12B text encoder is the memory cost; the int8 "lower-VRAM" files and LTX Desktop's free cloud text encoding both cut it.

Visual 5 · Prompt formula

The 6-part prompt LTX is trained to follow

From LTX's own prompting guide: one flowing paragraph, present tense, 4 to 8 sentences, under 200 words. No tag lists. Prompts written for Kling or Seedance should be rewritten into this shape.

1ShotScale and angle: medium close-up, wide, low angle
2SceneLight, colours, textures, one light logic
3ActionA sequence of verbs: pours, looks up, smiles
4CharacterAge, clothes, features, emotion shown by the body
5CameraThe move, and what we see after it
6AudioAmbience, music, dialogue in "quotes"
1A medium close-up shot of a young chai seller in a busy Lahore market at golden hour. 3He pours steaming tea from a height into a small glass, then looks up at the camera and smiles. 4He wears a faded blue kurta and a white cap, sleeves rolled up. 2Warm orange sunlight cuts through dust in the air, string lights blur behind him. 5The camera slowly pushes in toward his hands. 6We hear glass clinking and market chatter, and he says, "Garam chai, fresh, fresh!"
Show emotion with the body. "His jaw tightens, he looks away" works. "He is sad" gives the model nothing to animate.
Write the pauses. The model will not add a beat you did not write. "She pauses" or "a beat of silence" makes room.
Keep text out. Short on-screen words are better in 2.5 but still not guaranteed. Add titles and logos in your editor.
Simple physics wins. Chaotic motion (crowds fighting, liquids everywhere) brings smears. One clear action per shot.
turn any idea into an ltx prompt · paste into chatgpt, claude or gemini
Turn my video idea into a prompt for the LTX-2.5 video model.

My idea: [e.g. a barber shaving a customer in an old Lahore shop, for an Instagram ad]
Shot type I want: [single shot, or multi shot with 2 to 3 cuts]
Format: [9:16 vertical or 16:9]
Any spoken line: [e.g. "First cut is free." or none]

Write it the way LTX's own prompting guide asks:
1. One flowing paragraph in present tense, 4 to 8 sentences, under 200 words.
2. Cover in this order: the shot (scale and angle), the scene (light, colours, textures), the action as a sequence of verbs, the character (age, clothes, features, emotion shown through the body), the camera move, and the audio (ambience, music, dialogue in quotation marks).
3. For multi shot: name every cut ("the view cuts to..."), re-describe the person the same way after each cut, and say whether the audio continues or changes.
4. One light source logic for the whole scene. No on-screen text, logos or captions.

Then give me 2 variations: one with a different camera move, one with different lighting.
Do not add tags, bullet lists or shot numbers inside the prompts. Ask me first if my idea is missing something important.
It worked if you get one paragraph with no bullet points, a named camera move, and the dialogue in quotes, plus two variations.
Visual 6 · Multi-shot

How to get 2 to 4 shots in one generation

LTX-2.5's new trick from the reel: 2 to 4 connected shots in one clip, same person, same light, same voice. The yellow line is the playhead; watch what carries across each cut.

Shots
WideRooftop at dusk, he walks in with a kite
Close-upSame grey hoodie, tugs the string
Low angleOrange kite climbs the sky
Person
Same identifiers every shot: "young man in a grey hoodie"
Audio
Wind continues across the cut
Strings begin
  1. Name every cut in plain words: "The view cuts to...", "A hard cut jumps to...".
  2. Re-describe the shot after each cut: size, angle, who is in frame.
  3. Repeat the same identifiers for the same person, so the model keeps them.
  4. Say what the audio does at every cut: continues, drops, or changes.
  5. Stay single-shot for lip-synced dialogue in one framing, and for image-to-video from a first frame.
Worth saving

10 LTX-2.5 prompts that follow the formula

Written to LTX's guide: shot, scene, action, character, camera, audio. Swap the subject and keep the shape. Save them all as one text file with the button.

1 · Product hero shot (single shot)

Ads and client demos. Swap the product.

paste into ltx
A slow orbiting shot around a matte black wireless earbud case resting on a wet slate stone. The lid opens by itself and a soft white glow rises from inside as tiny water droplets slide down the stone. Cool blue rim light outlines the case against a dark studio background, with a faint reflection on the wet surface. The camera circles a quarter turn to the right and ends on a close-up of the open lid. A low electronic hum swells, then a crisp click as the lid locks open.

2 · Talking presenter (dialogue)

Explainers and faceless channels. Dialogue goes in quotes.

paste into ltx
A medium shot of a woman in her thirties sitting at a wooden desk in a bright home office, looking straight into the camera. She has shoulder length black hair, a mustard sweater and small gold earrings. Soft window light comes from the left, a plant and a bookshelf are softly blurred behind her. She leans forward slightly, raises one hand and says in a calm, friendly voice, "Here is the one setting most people never change." The camera stays static. Quiet room tone, a faint keyboard click in the background.

3 · Three-shot mini story (multi shot)

Shows off LTX-2.5 multi-shot. Name every cut and say what the audio does.

paste into ltx
A wide shot frames a rooftop in Karachi at dusk, the city glowing below, as a young man in a grey hoodie walks to the edge holding a kite. Soft wind and distant traffic fill the air. The view cuts to a medium close-up of his face, the same grey hoodie, eyes narrowed against the wind as he tugs the string; the wind continues across the cut. A hard cut jumps to a low angle shot of the orange kite climbing into the purple sky, the string snapping tight, and a gentle string melody begins over the wind.

4 · Food close-up (single shot)

Restaurant clients love this one.

paste into ltx
An extreme close-up of a fork pulling apart a stack of fluffy pancakes as warm honey drips down the sides in slow motion. Steam curls up from the soft center and a pat of butter slides off the top. Warm morning light from a window on the right, shallow depth of field, a white ceramic plate on a light oak table. The camera holds still, then racks focus from the honey to the butter. Soft cafe ambience, a quiet clink of cutlery.

5 · Real estate walkthrough (camera move)

Describe what the camera reveals after it moves.

paste into ltx
A smooth gimbal shot moves forward through the open front door of a modern apartment into a bright living room with floor to ceiling windows. Late afternoon sun lays long stripes of light across a light wood floor, a cream sofa and a low marble table. The camera glides past the sofa and turns slightly left to reveal a balcony with a city skyline beyond. Sheer curtains move gently in the breeze. Quiet ambient room tone and faint birdsong from outside.

6 · Cinematic B-roll for reels (single shot)

Generic B-roll you can cut into talking-head reels.

paste into ltx
A close-up tracking shot follows a person's hands typing quickly on a laptop in a dark room lit only by the screen. Lines of code scroll on the screen and reflect in a pair of glasses resting beside the keyboard. Cool blue light on the hands, deep shadows around them, a steaming mug at the edge of the frame. The camera slides slowly from left to right along the keyboard. Soft rapid typing sounds and a low synth pad.

7 · Animate your own photo (image-to-video)

Use with the Image-to-Video template or tool. Describe motion, not the photo.

paste into ltx
The person in the image slowly turns their head toward the camera and breaks into a relaxed smile. Their hair moves slightly in a light breeze and the background stays still. The camera pushes in gently from a medium shot to a close-up over the length of the clip. The light stays the same as in the image. Soft outdoor ambience with distant birds.

8 · Stylized animation (style word first)

Style words from the official list: claymation, stop-motion, 3D animation, hand-drawn.

paste into ltx
A claymation scene of a small orange fox in a knitted scarf building a snowman in a tiny snowy village. The fox rolls a snowball across the ground, it grows bigger, and the fox pats it into place with both paws, then steps back proudly. Felt textures, soft warm window lights in the clay houses behind, gentle snowfall. The camera holds a static wide shot at the fox's eye level. A soft music box melody and the crunch of snow.

9 · Product unboxing with voice (multi shot)

Two shots, one voice that carries across the cut.

paste into ltx
A top down shot shows two hands opening a plain brown box on a white table and lifting out a silver smartwatch. A calm male voice says, "This is the only watch I wear now." The view cuts to a close-up of the watch on a wrist, the screen lighting up with a simple clock face, the same calm voice continuing across the cut, "Seven days on one charge." Bright soft daylight, clean white background, gentle acoustic music under the voice.

10 · Nature establishing shot (single shot)

Openers and transitions.

paste into ltx
An aerial drone shot glides slowly over a misty mountain valley at sunrise, a winding river catching the first golden light. Pine forests cover the slopes and thin clouds drift between the peaks. The camera moves forward and tilts down slightly to reveal a small wooden bridge over the river. Soft warm light, cool blue shadows in the valley. Wind, distant birds and the rush of water.
It worked if the clip matches the camera move you wrote and the sound you described. If the motion smears, cut the action down to one clear movement.
Fine-tuning

LoRAs: add a style, or train your own

A LoRA is a small add-on file that teaches the model one style, camera move or character without retraining all of it.

01 · EasiestLTX Desktop recipesOne-click styles and moves: Claymation, Cozy Felt, Dolly In, Jib Up.
02 · Ready-madeOfficial LoRAsLightricks' LTX-2.5 collection on Hugging Face: LoRAs and IC-LoRAs (camera control, upscaling, HDR).
03 · In ComfyUILoraLoaderModelOnlyAdd it after the Model output, route it to both sampler stages. Strength from the LoRA's card, else 1.0.
04 · Your ownLTX Trainer80 GB VRAM recommended; low-VRAM config for 32 GB (RTX 5090). Rent an A100 at $1.79/hr.

Training is in the same repo as Route C (ltx-trainer). Start with the text-to-video LoRA config: it only needs short videos with captions. LTX also ships a /train-model agent you can run inside Claude Code that walks you through it. Most LoRAs trained on LTX-2.3 also run on 2.5, but test before client work, and a LoRA never works across different model families.

When it breaks

The 6 errors people hit first

You seeWhat it meansFix
401 / 403 on downloadYou have not accepted the gated model termsAccept on the LTX-2.5 Hugging Face page, log in with a Read token
Out of memory (OOM)Clip too big for your VRAM720p, 5 s, lower-VRAM files, close other apps, or Route D
Red nodes in ComfyUIMissing custom nodes or old ComfyUIInstall Missing Custom Nodes, update ComfyUI, full restart
"Coming soon to your device"LTX Desktop thinks your machine is below the floorMac: macOS 26 and 32 GB+. PC: more VRAM. Or Route D
Frame count errorFrames must be 1 + a multiple of 8Use 97 or 121 frames; sizes divisible by 32
Different result every runThe seed is on randomizeSet the seed to fixed once you like a result
stuck prompt · paste into chatgpt, claude or gemini
I'm trying to run the LTX-2.5 video model locally and I'm stuck. Help me fix it one step at a time.

My route: [LTX Desktop / ComfyUI template / official LTX-2 repo / Runpod cloud]
My computer: [Windows 11 / Linux / Mac], GPU: [e.g. RTX 4070 12 GB], system RAM: [e.g. 32 GB], free disk: [e.g. 150 GB]
What I clicked or ran: [paste the exact command or the menu steps]
The exact error: [paste the full error text or describe the red node]
Versions if I know them: [ComfyUI version, NVIDIA driver, Python]

Rules:
1. Tell me the most likely cause first, in plain words.
2. Give me one step at a time and wait for my reply before the next step.
3. If my hardware is too small for what I tried, say so and tell me the closest setting that fits (lower resolution, shorter clip, the lower-VRAM model files, or a cloud GPU).
4. Never ask me to paste an API key or Hugging Face token.
It worked if the AI names one likely cause and gives you a single step, then waits.
Every official link

Bookmark these

Malik Muzamil
The 1% AI Circle

Join our community

This guide gets LTX running. Inside the community you get every AI video setup in one place, new tools every week, and real help when a download or a red node breaks on your machine. Post your first LTX clip and we'll help you turn it into content that gets clients.

Malik Muzamil · Helping creators and freelancers make money with AI · Built AI courses and programs for top 1% companies

Join our community