How to Use NeonClipping: Free AI Video Clipping Tutorial
The verdict up front: NeonClipping turns your long videos into vertical, captioned 9:16 clips on your own computer, completely free — and this NeonClipping tutorial shows you how, step by step. Free has a precise meaning here:
- MIT open-source licence — no paid plan, no subscription, no credits.
- No account, and no upload — everything runs locally on your PC.
- What free does not mean: guaranteed-viral picks, zero setup, or a finished Mac app.
Launched on GitHub on 19 November 2026, NeonClipping picked up 1,272 stars in about two days — a full how-to guide appeared within 24 hours. The heat is simple to explain: it does the job creators pay OpusClip $15–29 a month for, free, on your own machine, without your footage ever leaving your PC.
One honest line to keep in mind: NeonClipping finds loud moments, not guaranteed-viral ones. It builds a strong first-pass shortlist — you stay the editor.
What NeonClipping is — and what it is not
NeonClipping is a free, local, open-source clipping engine. Paste a YouTube, Twitch or Kick link — or point it at your own MP4 — and it turns the long video into vertical 9:16 clips with burned-in captions and, with the AI add-on installed, a title and hook for each.
Just as important is what it is not:
- Not a cloud service. No server farm, no account system — it runs on your computer, on a local page only your machine can reach.
- Not a posting tool. It exports finished MP4 and subtitle files; uploading them is your job — nothing is published or scheduled for you.
- Not a full video editor. It finds, cuts and styles clips — not a timeline editor for building long videos.
- Not a virality machine. It ranks moments by loudness, chat spikes and similar signals; picking the winners is still your call.
- Not a licence to clip other people. Your own videos and streams, or content you have permission to clip — more on this below.
How NeonClipping finds the moments
The app labels its two modes Fast (~2-3 min) and Deep (slower, smarter):
- Fast mode reads the video’s word-timed captions and scores the audio — loudness jumps, raised voices, quiet-to-loud transitions — and clips snap to sentence boundaries, so cuts never land mid-word.
- Deep mode adds two visual signals — optical-flow motion and cut density — to the same scoring: slower, and better for action-heavy footage such as gameplay.
A moment mix setting steers what wins: Balanced, Action or Dialogue. With the free Ollama add-on (below), a local AI model reads the captions, scores each moment and writes its title and hook.
What you need before you start
- A Windows PC for the polished path — the installer adds Python (3.10+) and FFmpeg if they are missing.
- Or Mac / Linux via the project’s install.sh — be warned: the project itself says it is new and not yet tested on a real Mac.
- Optional but recommended: Ollama with the free llama3.2:3b model, for AI scores, titles and hooks; without it, clips are picked by audio and motion score alone.
- Optional: an NVIDIA graphics card. It speeds up encoding (NVENC) and speech-to-text (faster-whisper); a CPU fallback works, just slower.
- Your own video: a YouTube, Twitch or Kick link, or an MP4 file.
How to install NeonClipping on Windows
- Download the installer. Get NeonClipping-Setup.bat from the project website or GitHub Releases — official links at the end of this article, and nowhere else.
- Double-click it. It checks your PC and installs Python and FFmpeg if they are missing.
- Let it finish. It installs into %LOCALAPPDATA%\NeonClipping, adds a Desktop shortcut and opens the app.
- Open the app. It runs in your browser at 127.0.0.1:8770 — a local address, so the page exists only on your own computer. There is no sign-up — there is nothing to sign up to.
Mac or Linux uses the repo’s install.sh instead — the project’s own flagged-untested path.
NeonClipping tutorial: making your first clips, step by step
The app’s panels — Live clipping, Creators, My clips, Edit saved clips, Style & extras, AI parts, Find my facecam — organise one straight-line workflow:
- Paste your source: a YouTube, Twitch or Kick link, or your own MP4. Press Make clips.
- Set the shape of the clips. Set minimum and maximum clip length (30–80 seconds to start) and source quality: 1080p by default; 720p, 1440p, 4K available.
- Pick the mode and the mix. Fast (~2-3 min) for talk and podcasts; Deep (slower, smarter) for action and gameplay; mix: Balanced, Action or Dialogue.
- Watch the clips arrive. They appear one by one. (Developer-reported, not a benchmark: 97 seconds for an 85-minute stream on an RTX 2060 PC.)
- Choose a layout for each clip. There are 11: camera above or below gameplay, blurred fill, full-screen crop, gameplay only, four corner overlays, circle facecam, side-by-side. Find my facecam places your facecam preset.
- Style the captions. Styles: GTA Neon, Impact, Clean, Hormozi, MrBeast. Animations: highlight (word by word), pop-in, pop-in line, one by one, build up, full line. Templates: X post (dark and light), TikTok hook, Breaking news.
- Fix anything in My clips. In Edit clip: correct caption words, re-time captions, change layout or style — no regeneration needed.
- Export and post. Export the MP4s (1080 × 1920, captions burned in) plus an SRT file if you want one, then post them yourself.
Bonus — live clipping: paste a Twitch or YouTube Live link into the Live clipping panel, press Start, and clips are cut around chat spikes while the stream runs. The Creators panel can auto-clip a channel when it goes live — set it up for your own channel, or one whose owner has given you permission, never someone else’s without their say-so.
Free tools hub
Clip it free on your PC — then finish it free in the hub
NeonClipping makes the clips; our free tools hub finishes the job: thumbnails, a background remover, image resizers and hundreds more free tools, no sign-up.
Open the free tools hub
Ollama: the free add-on that writes the titles and hooks
Ollama is a separate free program that runs AI models locally — the one optional extra worth installing. Install it, pull the model with one command — ollama pull llama3.2:3b — and NeonClipping uses it to read the captions, score each moment and write every clip’s title and hook.
Without Ollama the tool still works — it finds and cuts clips by audio and motion score alone — but you lose the judgement layer: no AI scores, titles or hooks. One limit applies either way: the model reads text, the captions, so it judges what was said, not what was on screen.
Making GTA 6 clips with NeonClipping (your own footage only)
GTA 6 sessions are long, and the moments worth posting are buried inside them — exactly the problem this tool solves. Feed it your own gameplay recording or stream VOD, choose Deep mode so motion and cut signals help the audio score, and set the moment mix to Action. There is even a caption style built for the job: GTA Neon, and facecam streamers get layouts to match — camera above or below the gameplay, or a circle facecam.
The rule here is the rule for the whole tool: clip your own GTA 6 footage — your recordings, your streams, your VODs. Never download or re-clip another player’s footage without permission, and never use the Creators panel on anyone else’s channel.
Is NeonClipping really free?
Yes — and it is worth being exact, because “free” hides a lot here. NeonClipping is released under the MIT licence: no paid plan, no subscription, no credit system, no account. We tested it ourselves: the exported clips carry no watermark — the only thing the tool adds to your video is the captions you asked for.
What you “pay” with is your own computer: it does the processing, slower on a weaker machine. The optional AI layer costs nothing either — Ollama and llama3.2:3b are free and local. What free does not mean: cloud convenience, a guaranteed viral pick, or a polished Mac app.
Safety and privacy: your footage stays on your PC
- Nothing is uploaded. It runs a local server bound to 127.0.0.1 only, on port 8770; your videos, clips and captions never leave your computer.
- No account means no profile. No sign-up means no account holding your viewing or clipping habits.
- The code is public. The project is open source on GitHub — what it does is open to inspection.
- Permission is the boundary. The tool accepts any link you paste; the right to use that footage is yours to settle first.
- One known wrinkle: YouTube can show a “Sign in to confirm you’re not a bot” check when a video is fetched; the README has a section on it.
NeonClipping vs OpusClip: the honest comparison
| NeonClipping | OpusClip | |
|---|---|---|
| Price | Free — MIT open source, no paid tier | Subscription, $15–29/mo |
| Where it runs | Your own PC, locally | Their cloud servers |
| Your footage | Never uploaded | Uploaded to the service |
| Account | None | Account and subscription |
What the subscription buys is convenience: no install, no load on your PC, a polished cloud dashboard — at the price of a monthly fee and your footage on someone else’s server. NeonClipping is the opposite trade: zero cost and full privacy, in exchange for an install step and your own hardware doing the work.
The honest limits
- It finds loud moments, not guaranteed-viral ones. Volume, raised voices and chat spikes are proxies for excitement, not proof of it.
- The AI reads text, not pictures. A quiet visual moment can be missed entirely.
- No captions means slower work. If a video has no captions, the audio must be transcribed on your CPU.
- Windows is the polished path. The Mac and Linux installer is new and, in the project’s own words, untested on a real Mac.
- Live clipping has edges. The app must stay open, one stream at a time, and unattended recordings stop after 3 hours.
- Speed depends on your hardware. The timings quoted here are the developer’s reported figures from an RTX 2060 PC, not independent benchmarks.
Who should use NeonClipping — and who should skip it
Use it if you are a YouTube automation creator, streamer, podcaster or small studio sitting on long videos, want Shorts, Reels and TikToks from them without a monthly bill, and are happy to stay the editor. Sellers, note the angle: a product video or livestream becomes a week of Shorts — and a Short is a free ad.
Skip it if you are Mac-only and need a finished, tested app today; if you want clips auto-posted or scheduled; if your PC is too weak to stay on for processing; or if your plan involves clipping other creators’ content without permission.
Frequently asked questions
Is NeonClipping really free?
Yes: open-source software under the MIT licence — no paid plan, no subscription, no credit system, no account to create. The costs that remain are yours: the computer that does the work, and the optional free Ollama add-on for AI titles and hooks.
Does NeonClipping upload my videos to a server?
No. Everything runs on your own machine, through a local page only your computer can reach (127.0.0.1, port 8770); your footage is never uploaded to a cloud service — the main difference from subscription clippers.
Do I need a graphics card to use NeonClipping?
No. An NVIDIA card speeds up encoding and speech-to-text, but a CPU fallback exists and works — just slower, especially when a video has no captions and the audio must be transcribed.
Do I have to install Ollama?
No — it is optional. Without it, clips are still found and cut, chosen by the audio and motion score alone, but you get no AI scores, titles or hooks. Installing Ollama and the free llama3.2:3b model turns those on, locally.
Do the exported clips have a watermark?
No. We tested it ourselves: the exported clips carry no watermark, and there is no paid tier to remove one. You can add your own logo if you want branding.
Can I use NeonClipping on someone else’s videos or streams?
Only on your own videos and streams, or content you have permission to clip. The tool will accept a link, but the right to repost someone else’s footage is yours to have first — get permission, and use features like the Creators panel for your own channel.
Does NeonClipping post clips to TikTok or YouTube automatically?
No. It exports finished MP4 files (1080 × 1920) and subtitle files; you upload and post them yourself. Nothing is published or scheduled by the tool.
How long does it take to make clips from a long video?
The developer reports, on an RTX 2060 PC: about 97 seconds for an 85-minute stream, 136 seconds for a 3-hour Twitch VOD, 14 seconds for a 90-second file. Developer-reported figures, not independent benchmarks — a slower machine, or CPU transcription, takes longer.
Why did NeonClipping pick a loud moment instead of the best moment?
Because loud is what it measures: volume jumps, raised voices, quiet-to-loud transitions, chat spikes, the words in the captions. It finds loud moments, not guaranteed-viral ones, and a quiet visual moment can be missed because the AI reads text. Treat the clips as a first-pass shortlist; stay the editor.
What is the difference between NeonClipping and OpusClip?
NeonClipping is free, open-source software on your own computer: no account, subscription or upload. OpusClip is a cloud subscription at $15–29 a month, where you upload footage to their servers. NeonClipping asks more of your PC and your editing eye; OpusClip asks for a monthly fee.
Final verdict
NeonClipping deserves its first-week heat: a genuinely free, open-source clipper that respects your footage, your wallet and your judgement. It will not make a video go viral — nothing honest can promise that — but it turns the slowest part of a creator’s week — hunting for clip-worthy moments — into a two-minute shortlist. Install it, add Ollama, run it on your own footage, and stay the editor.
- GitHub repo (download and code): github.com/iamajmeena/neonclipping
- Project website: iamajmeena.github.io/neonclipping
- Guides: free English and Hindi PDFs in the repo’s docs folder.
More for creators and sellers
How this article was made
Written on 21 November 2026 from the project’s own GitHub repository and README: the features, free MIT terms, platform requirements and limits above all come from the project’s own documentation. Interface details — mode labels, moment mix, 11 layouts, caption styles and animations, templates, quality options and panel names — were checked against the app’s own code. Processing times are the developer’s reported figures from an RTX 2060 PC, labelled as such, not independent benchmarks. We also ran the tool ourselves for this article: a 70-second GTA 6 clip in Fast mode came back as two finished 1080 × 1920 clips with burned-in captions and no watermark. The finished-clip screenshot above is a frame from that test run, not an AI-generated image. Drafted with AI assistance; checked by an editor.
© 2026 VeloSpark Digital · Practical guides, checked against the sources themselves.

Written by Umar Hayat, founder of VeloSpark Digital and owner of Umar and Brothers (computers and mobile) in Toba Tek Singh, Punjab, Pakistan. He plans, builds and fact-checks every tool and guide on this site.
Related free tools: Background Remover · Image Compressor · Resize Image to KB · All free tools
Need an AI video ad, a logo or a design for your business? See our services.
Enjoying our free tools and guides? Add VeloSpark Digital as a preferred source on Google
