NodeCut analyzes your footage on your own GPU — cuts silences, removes filler words, selects the best camera angle. No uploads. Your files stay yours.
Free 7-day trial · No credit card · Works offline
Cloud tools process your clients' most personal moments on someone else's server. NodeCut runs entirely on your hardware — the RTX in your workstation or the M-chip in your MacBook. No upload queue. No bandwidth limit. No data terms to worry about.
Whisper large-v3 transcribes with word-level timestamps. Every pause, stutter, and filler word is located precisely before a single cut is made.
Every camera angle is scored every moment. The best shot wins — based on activity, face presence, sharpness, and emotion detection via MediaPipe.
Watch your timeline assemble in front of you, 30–60 seconds at a time. Pause, adjust, and resume from wherever the playhead sits.
No black box, no "done when done." NodeCut processes footage in 30–60 second chunks and applies cuts to your timeline immediately. You see everything as it happens — and can step in at any time.
Audio and visual analysis run in parallel. The only bottleneck is Whisper — everything else processes alongside it.
| Hardware | Processing time | Relative speed | |
|---|---|---|---|
Windows · RTX 3090 (CUDA) Your current setup |
~20–25 min | Fastest | |
| MacBook M4 | ~50 min | Fast | |
| MacBook M1 / M2 | ~70–80 min | ||
| Windows · CPU only (i7) | ~2.5–3 hrs |
Estimates for 5-hour wedding with 3 camera tracks. Whisper large-v3. Actual times vary with footage complexity.
No per-minute fees. No cloud processing charges. One license — unlimited footage, unlimited projects.
Be first in line for the $29 founder deal and beta access with real footage before public release.
No spam. Launch notification only. Unsubscribe anytime.
✓ You're on the list. We'll reach out before launch.