OpenCutAI now generates video from text — 9 AI models, 5 providers, built right into the editor

Hey everyone!

We just shipped an AI Video Generation hub directly into OpenCutAI. Generate video from a text prompt, preview it, and drop it on your timeline — all from one panel. No switching between browser tabs.

What's supported:

- 9 models across 5 providers — Runway Gen-3 Alpha, Pika 1.0, Kling v1.6 Pro, MiniMax Video-01, Stable Video Diffusion, Seedance 2.0, Stable Video XT, Luma Dream Machine, and CogVideoX (local, free, no API key)

- 3 generation modes — Text-to-Video, Image-to-Video, Video-to-Video

- 6 aspect ratios — 16:9, 9:16, 1:1, 4:3, 3:4, 21:9

- Prompt enhancement — write a rough idea, click Enhance, and the LLM expands it into a detailed cinematic prompt

- One-click add to timeline — preview the result, then drop it straight into your project as an editable clip

- Generation history — all your previous generations are saved so you can reuse them without re-running

How it works:

1. Pick a provider and model (or filter by provider)

2. Write your prompt

3. Choose aspect ratio and duration

4. Hit Generate

5. Preview the result with playback controls

6. Click "Add to Timeline"

API keys are managed in Settings. One key per provider — Replicate ($5 free credits), Stability AI, Luma AI, Seedance. Or skip all of them and use local CogVideoX for free.

Also shipped in this update:

- Smart Reframe — detects faces and auto-generates keyframes to keep subjects centered when converting landscape to vertical (9:16)

- Chroma Key — green screen / blue screen effect with WebGL shader, 5 presets, tolerance, and spill suppression

- Motion Tracking — client-side object tracking that generates position keyframes. Runs entirely in the browser

Everything is open source, MIT licensed. Self-host on a $20/mo server or run locally.

What models or providers would you want added next?