can i fork it?
open sourceApache-2.0Local or hosted★ 1.9kactive jul 2026

LTX Desktop

A real local video editor with text and image generation built in; your GPU replaces the credit card.

Technical Profile & Specs

verified 2026-08-10

Cost
$0 / forever (Open Source)
License
Apache-2.0
Hosting Model
Local or hosted (your machine or their cloud)
Platforms
desktopone-click install
Installation
Download the Windows .exe or macOS .dmg from the project's GitHub Releases page
Operational Setup
Serious local generation currently needs a high-end NVIDIA Windows GPU and very large model downloads; macOS generation relies on an API key instead of local inference
Where Your Data Lives
Projects, settings, logs, downloaded models, source assets, and renders live in the platform-specific LTXDesktop application-data folder; API mode uploads the material needed for the requested generation
The Catch / Trade-off
It substitutes LTX models for Hailuo's MiniMax model, so the characteristic motion and quality differ even before the user confronts the demanding local GPU requirement or API bill.

Replaces 6 paid apps

up to $15/mo ($180/yr) in subscriptions

LTX Desktop FAQ

Common questions

Is LTX Desktop free and open source?+

LTX Desktop is an open-source tool licensed under Apache-2.0 that runs on local or hosted. A real local video editor with text and image generation built in; your GPU replaces the credit card.

What paid software does LTX Desktop replace?+

LTX Desktop serves as a free replacement for 6 paid SaaS applications: Hailuo AI, Kaiber, Kling AI, Luma Dream Machine, Pika, Runway, saving up to $15/month ($180/year) in subscription bills.

How do you install and set up LTX Desktop?+

Installation: Download the Windows .exe or macOS .dmg from the project's GitHub Releases page. Setup notes: Serious local generation currently needs a high-end NVIDIA Windows GPU and very large model downloads; macOS generation relies on an API key instead of local inference

What is the catch or trade-off with LTX Desktop?+

It substitutes LTX models for Hailuo's MiniMax model, so the characteristic motion and quality differ even before the user confronts the demanding local GPU requirement or API bill.