Roadmap
OpenCut AI is a fork of OpenCut with AI wrapped around it. Here's what's been built and what's next. (last updated: March 22, 2026)
- 1
OpenCut — the foundation
OpenCut is the open-source video editor this project is forked from. It provides the core editor — multi-track timeline, real-time preview, text/sticker/effect tracks, keyboard shortcuts, and browser-based storage.
Upstream - 2
Fork & AI integration
OpenCut AI is a fork that wraps AI capabilities around the core editor. The goal: make video editing accessible to non-editors by letting them edit videos through text, voice, and AI commands — all running locally.
Completed - 3
AI transcription & text-based editing
Backend Whisper service for speech-to-text with word-level timestamps. Transcript panel with live word highlighting, auto-scroll, segment deletion, word-level cuts, and drag-to-reorder. Video auto-splits at segment boundaries.
Completed - 4
Voice cloning & TTS
Coqui XTTS v2 for multilingual voice generation with voice cloning. Upload a voice sample to clone any voice. Generate voiceovers per-segment from the transcript. Background task tracking in the UI.
Completed - 5
Subtitles & multilingual support
One-click subtitle generation from transcript. Auto-translation of transcript to 12+ languages for multilingual subtitle tracks. Word-level karaoke highlighting.
Completed - 6
Filters, adjustments & effects
WebGL color-adjust shader with 12 filter presets. Adjustment panel with brightness, contrast, saturation, temperature, and vignette sliders. Effects applied as timeline tracks.
Completed - 7
Audio separation & volume control
Auto-separate audio from video into its own track. Per-clip volume control with draggable dB line. Volume changes via Web Audio GainNode.
Completed - 8
Overlays & compositing
Picture-in-picture presets, split screen, compositing presets (ghost overlay, dark overlay, light leak). Per-element opacity, 17 blend modes, and transform controls.
Completed - 9
AI Studio & fact-checking
AI chat for brainstorming, script writing, and content planning. Fact-check panel — extracts claims from transcript and verifies them via LLM.
Completed - 10
Quick actions & filler removal
Quick actions bar with one-click filler word detection, silence detection, subtitle toggle, podcast subs, clip finder, and fact-check.
Completed - 11
Self-hosting & infrastructure
Full Docker Compose setup with 10+ services (web, AI backend, Whisper, TTS, image, speaker, face, Ollama, Postgres, Redis). Health monitoring via proxied endpoint. RAM/GPU status in header.
Completed - 12
Podcast Clip Studio
AI-powered clip finder that scores moments by viral potential. Word-pop karaoke subtitles (Hormozi style). AI question cards with transparent overlays. Keyword highlighting via LLM. Emotion detection with speechbrain.
Completed - 13
Multi-speaker detection
pyannote speaker-diarization-3.1 service with silence-based fallback. Editable speaker names in transcript. Auto-cut at speaker boundaries. Speaker legend bar with color-coded labels.
Completed - 14
Auto-reframe & face tracking
MediaPipe face detection service. Auto-crop 16:9 to 9:16 for TikTok/Shorts. Speaker-aware face centering based on diarization positions. Smooth pan transitions.
Completed - 15
Brand Kit
Full brand identity with colors, fonts, logo, social handles, CTA. Intro/outro cards, lower thirds, persistent watermarks. Font picker with Google Fonts + local upload. Data persisted to localStorage.
Completed - 16
Speed control & auto-save
0.1x to 4x playback speed with 8 presets + slider. Audio pitch adjustment. Auto-save with recovery prompt. Save status indicator in editor header.
Completed - 17
Export
MP4 (H.264) and WebM (VP9) export with quality presets. Platform presets for YouTube, TikTok, Instagram. Progress bar with cancel support.
Completed - 18
TurboQuant KV Cache Compression
Google Research-based PolarQuant + QJL algorithms compress KV cache up to 6x. Run 7B models on 8GB RAM. Full AI stack reduced from 35GB to 15GB. Performance tiers: Lite, Standard, Pro.
Completed - 19
Advanced export & rendering
Subtitle burn-in during export. Background export. Batch export for multiple formats. Custom resolution and bitrate controls. GPU-accelerated rendering.
Planned - 20
Transitions & smart suggestions
Fade, dissolve, slide transitions between clips. Post-upload workflow wizard. One-click auto-edit. Audio ducking for background music.
Planned - 21
More AI models
Model download manager. Support for StyleTTS 2, Bark, Piper, Fish Speech, Kokoro for TTS. Switchable Whisper model sizes. Cloud API fallbacks as optional alternatives.
Planned - 22
Native app (mobile/desktop)
Native OpenCut AI apps for Mac, Windows, Linux, and iOS/Android.
Not started
Want to help?
OpenCut AI is open source and built by the community. Every contribution, no matter how small, helps us build the best free video editor possible.