About vid3k
vid3k is an AI-powered short-form video production platform. Give it a script and a reference image, and it delivers a finished vertical video — fully edited, captioned, and ready to post.
How it works
Every video passes through an 8-stage automated pipeline. You write the script, choose a style preset (or fine-tune every parameter), and hit generate. Minutes later you have a polished 9:16 vertical ad.
- Queued — Your job enters the pipeline.
- Generate — AI creates video clips from your script and reference image using native audio synthesis for dialogue.
- Analyze — Whisper transcription extracts precise word-level timing.
- Beat Sync — Cuts snap to the musical beat for a professional rhythm.
- Captions— Word-level karaoke-style captions highlight each word as it's spoken.
- Transitions — Fade, dissolve, wipe, slide, and zoom transitions with timed SFX hits.
- Music — Background music is auto-ducked during speech and mixed to taste.
- Color — Color grading and LUT application in a single pass.
Technology
vid3k combines several AI and media-processing technologies into a single pipeline:
- Video generation — Powered by Atlas Cloud API (Veo 3.1 Lite) for high-quality AI-generated video clips with native audio.
- Speech analysis — OpenAI Whisper provides word-level transcription for precise caption timing.
- Post-production — FFmpeg-based pipeline handles beat detection, transitions, audio mixing, color grading, and caption rendering.
- Infrastructure — Serverless processing, edge-deployed frontend, and zero-egress media storage keep costs low and latency minimal.
Built for creators
Whether you're producing product launches, testimonials, tutorials, or promo ads — vid3k handles the editing so you can focus on the message. One-click presets get you started fast, and granular controls let you dial in every detail when you need to.