Is there a free open source AI video generator?

Yes, in both senses of the phrase. For text-to-video: Mochi 1, Open-Sora and HunyuanVideo, all GPU-hungry. For clipping your own footage into shorts: OpenShorts, MIT-licensed, free self-hosted or hosted from $12/month.

Can open source AI generate videos from text?

Yes: Mochi 1, Open-Sora and HunyuanVideo publish open weights. OpenShorts is not one of them; it turns long real footage into short vertical clips.

Home Open source AI video generator

An open source AI video generator, in the sense that matters for creators

By the OpenShorts team· Published 2026-08-04· Updated 2026-08-04

TL;DR

"AI video generator" names two different products. Text-to-video models invent new footage from a written prompt. Clip generators produce short videos from long footage you already have. Confusing the two wastes an afternoon.

For text-to-video there are real open source options, including Genmo’s Mochi 1, Open-Sora and HunyuanVideo, all of which need a serious GPU.

For turning your own recordings into vertical shorts, OpenShorts is MIT-licensed and self-hosted: transcription, AI moment scoring, face-tracked 9:16 reframing and burned-in subtitles, free on your own machine or hosted from $12/month.

Which "AI video generator" are you looking for?

If you type this query wanting a model that produces footage from a text prompt, you want a text-to-video model. If you have a podcast, webinar, stream or interview recording and want short vertical videos out of it, you want a clip generator. The two share almost no technology and no workflow. This page covers both honestly and goes deep on the second, because that is what OpenShorts is.

Open source text-to-video, briefly

As of August 2026 the notable open-weight text-to-video models include Genmo's Mochi 1 (Apache 2.0), Open-Sora, Tencent's HunyuanVideo and Alibaba's Wan family. They genuinely generate novel footage, and they need data-center or high-end consumer GPUs to run at usable speed. If that is your goal, start with those projects; OpenShorts will not do it.

Generating videos from footage you already have

OpenShorts is an open source AI clip generator that turns long videos (podcasts, webinars, livestreams, interviews) into vertical 9:16 clips for TikTok, Instagram Reels and YouTube Shorts.

faster-whisper transcribes the video with word-level timestamps, PySceneDetect finds the scene boundaries, and Google Gemini 3.0 Flash scores the transcript to pick the 3 to 15 strongest moments of 15 to 60 seconds each. Each moment is then cut with FFmpeg and reframed to 9:16 with MediaPipe face tracking.

OpenShorts comes in two editions and they are priced very differently, so it is worth being precise. OpenShorts (self-hosted) is free and open source under the MIT licence: Free and open source under MIT. You run it with Docker on your own machine and bring your own API keys. No watermark, no usage cap, no subscription. What it costs you is hardware and time: on a typical CPU an 8-minute video takes 5 to 8 minutes to process. OpenShorts Cloud is the hosted service: The same software with the compute and API keys covered. An NVIDIA GPU clips that same 8-minute video in about 50 seconds. Free plan is 20 minutes a month with a watermark and no credit card. Paid plans start at $12/month for 100 minutes with no watermark, up to $59/month.

Other open source clip generators, compared honestly

Checked 2026-08-04 on GitHub. Star counts move; positioning rarely does.

OpenShorts is not the only open source project in this space, and pretending otherwise would not survive one GitHub search. The notable neighbours:

AI-Youtube-Shorts-Generator: the most-starred repo in the category, with a leaner scope built around highlight extraction and cropping.
supoclip and clippyme: smaller projects covering transcription-driven clipping, the latter also using Gemini for moment selection.
MoneyPrinterTurbo: generates videos from text plus stock footage, which is a different job than clipping your own recordings.

Where OpenShorts differs from all of them is surface area: a web dashboard, a REST API with keys, completion webhooks, an MCP server for agents, split-screen and screencast layouts for two-person and screen-share footage, dubbing into 30+ languages, and direct publishing to TikTok, Instagram Reels and YouTube Shorts. If you want a small script you can read in an hour, the smaller repos are a better fit, and that is a real recommendation rather than false modesty.

Common questions

Is there a free open source AI video generator?: Yes, in both senses. For text-to-video, Genmo’s Mochi 1, Open-Sora and HunyuanVideo publish open weights and need a powerful GPU. For making clips from your own footage, OpenShorts is MIT-licensed and runs with Docker on an ordinary machine: free self-hosted with no watermark, or hosted from $12/month.
Can open source AI generate videos from text?: Yes. Mochi 1 (Apache 2.0), Open-Sora and HunyuanVideo generate footage from prompts. Expect to need a high-end GPU, and expect quality below the closed frontier models. OpenShorts is not a text-to-video tool; it turns long real footage into short vertical clips.
What is the best open source AI video generator for shorts?: For turning long recordings into publishable vertical shorts with subtitles, OpenShorts covers the widest pipeline: AI moment scoring, face-tracked reframing, split-screen layouts, dubbing and direct social publishing, MIT-licensed. Simpler repos like AI-Youtube-Shorts-Generator cover a leaner version of the same job with less to configure.

Sources

Open-weight text-to-video model landscape checked 2026-08-04 on the respective GitHub repositories.
OpenShorts source at github.com/mutonby/openshorts.

Related comparisons

How a long video becomes a vertical clipThe full pipeline, stage by stage.Turn a podcast into shorts without cropping out half the conversationTwo-speaker episodes without cropping anyone out.A YouTube to Shorts converter that starts from the linkPaste a link, get 9:16 clips with subtitles.