HunyuanVideo

A large open-weights text-to-video model from Tencent. Genuinely good output, and a genuinely large hardware bill.

Self-hosted
Active
Unknown adult content
Open source

At a glance

Where it runs: Self-hostedYou run it on your own machine. Nothing leaves it unless you send it.
Adult content: UnknownWe have not found a clear answer. We would rather say so.
Trains on your content: NoYour content is not used to train models.
Status: ActiveMaintained and working today.
Account needed: No
Licence: Tencent Hunyuan Community License
Cost: Free weights under a community licence with use restrictions.
What it keeps: Local only.

What it takes to run

Graphics memory: 16 GB VRAM, 48 GB recommended
System memory: 32 GB, 64 GB recommended
Disk: about 80 GB
Runs on: Linux

Find what fits your machine on the tools page — pick your hardware from the first menu.

One of the first open-weights video models whose output stood up next to the closed services. The weights are downloadable and it runs locally, which for video is a much bigger deal than it sounds.

The requirement is the catch: a full-quality generation wants far more VRAM than a consumer card has, and the community quantisations that bring it down to 12 or 16 GB cost quality and time.

Worth knowing before you start

Very heavy Full quality is a datacentre card; consumer use means quantised community builds.
Community licence, not open source The licence carries use restrictions and a territorial exclusion.

What it does

Open weights
ComfyUI support
Text and image to video

How it relates to other tools

requires ComfyUI — the usual way to run it on a desktop

Signs of life

Checked automatically. These are the only figures on this page a machine wrote, and they say when they were taken.

Last answered: Yes, 3 hours ago Its website responded when we asked.
Stars on GitHub: 12,527 1,331 forks.
Last commit: 78 days ago2026-06-29 09:33 UTC
Archived copy: Wayback Machine A snapshot, in case this one stops answering.