OpenCut AIGive your idea a storyboard.
Plan shots and generate media with AI. Place them sequentially on the timeline, with placeholders, retries and optional frame references.
AI STORYBOARDOPEN SOURCE. OPEN POSSIBILITIES.
Turn ideas
into frames.
From your first recording to your final frame.
Bring AI storyboards, media, sound and subtitles into one creative workspace.
macOS · Windows · Linux / Local editing. No subscription.
OPENCUT AI PROJECT / 001
A story begins.Every story deserves to be seen.
00:00:12:08◉ SEQUENCE 0116:9 4K
00:00 00:0500:1000:15
01 / IDEA02 / CONTINUE03 / FINISH
Every story deserves to be seen.
∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿∿
Concept illustration · Download the desktop editor
Your story. Your pace. Your workspace.01 — RECORD 02 — GENERATE 03 — EDIT 04 — EXPORT
FROM IDEA TO FINAL FRAME
Less switching.
More creating.
Capture the story on screen.
Screen recording, separate system audio and compact controls. Pause and resume to turn your process into editable footage.
SCREEN RECORDINGMake every frame your own.
Customize subtitle fonts, colors and shadows, image overlays, effects, transitions and sounds. Preview changes and include them in your export.
CRAFT EVERY FRAMEAI / UNDER THE HOOD
Models, keys and generation limits.
AI currently connects through ZenMux. Base URL: https://zenmux.ai/api/v1. Save your ZenMux API Key in Settings → ZenMux AI. Separate OpenAI, Google or MiniMax keys are not required. ZenMux bills model calls against your account; local editing requires no subscription.
| Use | Preset model IDs | API / limits |
|---|---|---|
| Video | minimax/minimax-h3-max | /videos · 5–15 seconds per shot · 480p / 768p (default) |
| Video | google/gemini-omni-1.1-flash-preview | /interactions · Duration and aspect ratio are creative preferences; exact output is not guaranteed |
| Image | openai/gpt-image-2.5-flareopenai/gpt-image-2.5-sunburstgoogle/gemini-3.1-flash-image | Image generation / Google generateContent route; output dimensions depend on the model |
| Storyboard and edit planning | openai/gpt-6.1-solanthropic/claude-sonnet-5.5 | /chat/completions · Editable storyboards and edit plans |
These are preset IDs in the current code, not a guarantee of availability. Check the ZenMux catalog, account access and balance. Model lists are editable. Storyboard shots span 1–30 seconds and split at configured model limits. Standard video requests validate 1–120 seconds, but providers may impose lower limits. There is no universal resolution guarantee for other models. AI output resolution is separate from project and export resolution.
Each shot can optionally pass its last frame as the next shot’s first-frame reference. The route must support image references; visual or character continuity is not guaranteed. Prompts and reference media are sent to the provider. Keys stay locally and are excluded from project exports.
FIND YOUR WORKFLOW
Why choose OpenCut AI?
| Product | Main focus | The OpenCut AI approach |
|---|---|---|
| CapCut | Editing, templates and publishing for a broad creator audience | For creators who want open source, their own API key and editable AI storyboards |
| DaVinci Resolve | Professional editing, color grading, audio post-production and visual effects | A focused recording → AI storyboard → timeline workflow. Resolve remains a better fit for professional grading and VFX. |
| OpenCut Classic | The open-source editor foundation inherited by this project | Adds AI storyboards, placeholders and retries, optional last-frame references, screen recording and system-audio workflows |
This compares workflow positioning, not benchmark results or exhaustive feature sets. OpenCut AI builds on OpenCut Classic and retains its MIT license and notices.
An open starting point.
Edit freely.
Express yourself.
Build with us on GitHub ↗Version 0.1.9 adds balance and budget settings using a separate Management API Key, plus recording recovery, a task journal and selectable continuity frames. Version 0.1.11 adds direct subtitle content editing with immediate previews, undoable saves and matching exports, and includes the microphone permission fix. Setup and practical limits →
YOUR NEXT STORY STARTS HERE
Choose your workspace.
Local editing. Open source. AI generation requires your own service account and API key.
macOS
Apple Silicon · DMG
Developer ID signed · Apple notarized
Windows
x64 · Setup installer
Unsigned: Windows SmartScreen may warn you. Verify the release source and SHA256 checksum first.
Linux
x64 · Debian / Ubuntu / Fedora
Screen capture support varies by desktop environment.
Find the latest version and all installers on GitHub Releases ↗