AI Clips · Pro · Windows

AI short video generator

Turn your long videos into captioned vertical shorts automatically — transcription and rendering run on your own PC. Hand VidFabs a podcast, stream replay, lecture or gameplay video and get back ready-to-post clips.

Download for Windows See Pro plans
Finished AI Clips job: captioned highlights ready to preview

How it works

  1. Queue your videos

    Add one long video or up to 20 at once; they're processed one after another.

  2. Download the speech model (first run only)

    The speech-recognition model downloads once with a progress bar and is cached on your PC.

  3. Transcribe locally

    Multilingual Whisper models turn speech into timestamped text on your own machine — the audio never leaves it.

  4. Pick the highlights

    The transcript goes to the LLM you configured, which picks up to 10 highlights per video.

  5. Render the shorts

    Each highlight is cut, reframed and rendered with captions — burned into the video or saved as a separate .ass file, your choice per video. Preview them, read the captions and open the output folder.

AI Clips pipeline with queued videos and per-stage progress

Output framing

  • Vertical 9:16 — crop: fills the vertical frame, made for short-video platforms.
  • Vertical 9:16 — full frame: keeps the whole picture with blurred bars above and below.
  • Original 16:9 — full frame: classic landscape output.

One-time setup

  • Speech model — pick a recognition model in Settings → AI. Larger models are more accurate, and Pro includes the most accurate one.
  • LLM API — your own OpenAI-compatible key, with provider presets or a custom base URL (Azure, local servers, proxies). The key is stored only on your PC.
Settings → AI: speech model and LLM API configuration

Local, not uploaded

Cloud clipping tools make you upload the whole video first and often charge by the minute or by credits. VidFabs transcribes and renders on your PC; the only thing sent out is the transcript text, to the LLM endpoint you chose, billed to your own key.

Try it free with a sample video

No API key yet? On the free plan you can run the full pipeline on a NASA sample video — real model download, local transcription and rendering, with pre-picked highlights. It needs an internet connection to download the sample.

Fine-tune in the Advanced editor

Import the generated clips into the Advanced editor (Media Bin → Import media) to fine-tune them frame by frame.

See the frame-by-frame editor →

Frequently asked questions

Do I need an API key?
Yes — AI Clips is a Pro feature that uses your own OpenAI-compatible API key to pick highlights, and VidFabs never bills you for model usage. Without a key you can still try the full pipeline on the free sample video (it needs an internet connection to download the sample).
Is my video uploaded?
No. Transcription and rendering run on your PC. Only the transcript text is sent to the LLM endpoint you configure.
How many clips do I get?
Up to 10 highlights per video, and you can queue up to 20 videos at a time.
Which aspect ratios can it output?
Vertical 9:16 (cropped to fill, or the full frame with blurred bars) or the original 16:9.
Which languages does it support?
Captions come out in the video's spoken language, auto-detected by multilingual Whisper models. Accuracy varies by language and model size.

Turn your next long video into shorts

Download free for Windows — upgrade to Pro for AI Clips and the Advanced editor.

Download for Windows See Pro plans
Read the full AI Clips guide →