Local video generation
Windows desktop app, a compatible modern NVIDIA GPU and at least 16 GB of VRAM. Compatible models include RTX 3090, RTX 4090 and RTX 5090. The minimum refers to GPU memory, not system RAM.
Local AI video generator
Create faceless videos on your Windows PC with a compatible NVIDIA GPU and at least 16 GB of VRAM. Or rent a GPU from the app on Windows or Mac. Aificient Studio brings stories, reusable characters and animated promotional posters into one workspace.
Local rendering: Windows + compatible NVIDIA GPU · 16 GB+ VRAM
Runs on your machine
Choose local rendering on a compatible Windows PC, rent a GPU from the Windows or Mac desktop app, or send any scene to Aificient's generation cloud, available 24/7. The web app also lets you create from a computer, phone or tablet.
Windows desktop app, a compatible modern NVIDIA GPU and at least 16 GB of VRAM. Compatible models include RTX 3090, RTX 4090 and RTX 5090. The minimum refers to GPU memory, not system RAM.
The Windows and macOS desktop apps include a GPU marketplace. Choose an RTX 4090, RTX 5090 or H100 and rent it with one click, billed by the second: from $0.40 an hour for an RTX 4090, $0.70 for an RTX 5090 and $2 for an H100. You do not need a powerful local GPU to use a rented one.



Requirements
Using Aificient and rendering video on your own GPU have different requirements. Choose the setup that fits your device.
Windows desktop app, a compatible modern NVIDIA GPU and at least 16 GB of VRAM. Compatible models include RTX 3090, RTX 4090 and RTX 5090. The minimum refers to GPU memory, not system RAM.
The Windows and macOS desktop apps include a GPU marketplace. Choose an RTX 4090, RTX 5090 or H100 and rent it with one click, billed by the second: from $0.40 an hour for an RTX 4090, $0.70 for an RTX 5090 and $2 for an H100. You do not need a powerful local GPU to use a rented one.
Use the web app on a computer, phone or tablet. Video assembly is available through an API with a small usage cost. The desktop apps can perform assembly locally, avoiding that API assembly cost. Generation itself runs on Aificient's cloud, available 24/7, and any scene can be sent there from the desktop apps too.
Proof, not stock
Every clip here was generated frame by frame inside Aificient Studio — rendered for the story, never pulled off a stock shelf. One pipeline, completely different worlds.
Reference times
Reference generation times from the Aificient team, using Lite mode. Compare the length of the finished clip with the time spent generating it.
Lite mode · 1080p · approximate generation time
Pro mode takes approximately 1.5× as long as Lite. These are indicative times per clip, not a promise for every scene or a total for a finished, multi-scene video. Actual times vary with settings and the scene.
Render → publish
Finish your video, prepare the post and choose when to publish from the desktop app. Keep creation and your social publishing workflow in the same workspace.
Connect the social accounts you want to use from the desktop app.
An AI caption in your tone, plus a real trending sound pulled live from inside your own TikTok session.
Push the finished 9:16 video live, or queue it for later — no re-upload, no second tool.
FAQ
The short version — follow any answer through to the full story.
Local GPU video generation currently requires Windows and a compatible NVIDIA GPU with at least 16 GB of VRAM. On Mac, use the desktop app with a rented GPU, or use the web app. The Mac desktop app can still perform tasks such as video assembly locally.
Download the desktop appCompatible modern NVIDIA GPUs with at least 16 GB of VRAM, including RTX 3090, RTX 4090 and RTX 5090. VRAM is the GPU's own memory. Without a suitable card, you can rent an RTX 4090, RTX 5090 or H100 from the desktop app.
Download the desktop appSignup includes free credits and a trial. After the trial, continued use requires a subscription, even if you render locally. Standard starts at €14.99 per month with 90 credits and 10 GiB of storage. Rented GPUs are billed by the second at the rate shown in the app.
See plans and pricingThe team's reference times put Pro at approximately 1.5 times the duration of Lite. For example, a 5-second 1080p clip takes about 110 seconds in Lite on an H100. Settings and scene content affect the actual time.
See the render timesYes. The web app works on computers, phones and tablets. Local NVIDIA rendering is a Windows desktop feature; the web app performs video assembly through an API with a small usage cost.
Open the web appLocal rendering describes where the video is generated. Aificient also includes online services, including the web app, cloud generation and GPU rental. Desktop video assembly can run locally; do not treat local rendering as a fully offline workflow.
See the requirementsTry it for free — post your first original short tonight.