Clipify
●Open-source · MIT Licensed · No cloud required

Turn any video into viral clips,
in one command.

Feed Clipify a long-form video or YouTube URL — it transcribes with Whisper, scores the best moments, cuts with ffmpeg (zero re-encode), and exports multi-ratio versions with per-clip captions. Fully automatically.

View on GitHubHow it works
GitHub StarsGitHub ForksMIT LicenseCI passing
bash — ~/projects

The problem

Manual clipping is a grind.
It shouldn't be.

Download the recording. Watch the whole thing. Find the good moments. Trim each clip. Add captions manually. Export three times — one for each platform ratio. Repeat for every video, every week.

Clipify collapses that entire workflow — download → watch → cut → caption → reformat ×3 — into a single command that runs while you do something else.

Before

✕Download file
✕Watch full video
✕Find moments manually
✕Cut each clip
✕Add captions
✕Export 9:16
✕Export 16:9
✕Export 1:1

After

✓python clipify.py --video … --clips 10
Everything else happens automatically

How it works

One command. Seven steps.
All automated.

The full pipeline — exactly as it runs in the repo, no steps invented.

01

Input

Local video file, YouTube URL, or a watched folder

→
02

Audio extraction

ffmpeg strips and normalizes audio without touching the video stream

→
03

Whisper transcription

Segment-level word timestamps from OpenAI Whisper — accurate, local, or cloud

→
04

Moment scoring

Audio energy peaks + hook signals + sentiment — each moment ranked automatically

→
05

Clip extraction

ffmpeg cuts precisely at scored timestamps — zero re-encode, lossless quality

→
06

Multi-ratio export

9:16 (Shorts/TikTok), 16:9 (YouTube), 1:1 (Instagram) — all generated in one pass

→
07

Captions

VTT caption file generated per clip, synced to word timestamps from Whisper

AI flexibility

Bring your own model.
Or run fully offline.

Clipify supports 7 AI providers — OpenRouter, Groq, DeepSeek, OpenAI, Anthropic, Gemini, and xAI — with automatic fallback and selection. Point it at whichever provider you already have a key for, or use none at all.

Pass --ai local and everything runs on your machine — no API key, no network call, no data leaving your system.

FEATUREAnti-Hallucination Mode

Constrains all generated outputs to transcript-grounded content only. No invented summaries or fabricated quotes — every title, caption, and hook is anchored to what was actually said. Essential for news clips, product demos, or any compliance-sensitive content.

// supported providers

OpenRouter
Groq
DeepSeek
OpenAI
Anthropic
Gemini
xAI
localzero dependencies

// automatic fallback & selection when a provider is unavailable

Why Clipify

Three things that actually matter.

Open source

No lock-in, ever.

MIT licensed. Inspect every module, extend any step, swap any AI provider through a documented interface. The entire pipeline is yours to fork, modify, and self-host.

→MIT license
→Every module is swappable
→No proprietary runtime
Privacy-first

Run entirely offline.

Pass --ai local and nothing leaves your machine. No video data uploaded, no API key required, no usage telemetry. Use cloud providers only when you explicitly choose to.

→--ai local mode
→Zero API dependency
→Data stays on your disk
Built for speed

Zero re-encode. One command.

Clips are cut at ffmpeg keyframe precision with zero re-encode — original quality preserved, extraction is near-instant. Full pipeline from raw input to platform-ready exports in one pass.

→Zero re-encode
→One command, all ratios
→Captions generated inline
Roadmap — not shipped yet

Where we're headed.

None of the following is built yet. This is the vision for what Clipify is growing toward — a full content operation, not just a clipping tool.

Hosted version

coming soon

No local setup required — upload straight from the browser, no CLI needed.

Automatic multi-platform publishing

coming soon

Schedule and post directly to YouTube, TikTok, Instagram, and LinkedIn instead of just exporting files.

Specialized agent team

coming soon

Dedicated agents for titles/thumbnails, SEO, per-channel copy (descriptions, hashtags, social posts), and community handling — each doing one job well.

Transparent scoring breakdown

coming soon

Every clip's score decomposed: hook strength, clarity, emotional impact, and more — so you see why a clip ranked highly, not just a single number.

Co-editing with notes

coming soon

Leave a note on any generated clip, title, or caption and have it regenerated with that guidance — no manual re-editing.

Memory that compounds

coming soon

The system learns which of your clips actually performed well and uses that to improve future suggestions — not the same generic heuristics forever.

Autonomy dial

coming soon

From 'review everything before it posts' to 'fully autonomous', with brand guardrails you set once.

Platform reliability layer

coming soon

Health monitoring, exponential backoff, and graceful handling of rate limits and API quirks that come with posting automatically at scale.

Team & enterprise features

coming soon

Multi-seat workspaces, approval chains, and audit logs for agencies and media teams managing multiple creators or brands.

// want to shape the roadmap? open an issue or start a discussion on GitHub.

⭐

It's open source.
Come build it with us.

Star the repo, try the CLI, open an issue, or submit a PR. Real working code, public commits, CI passing — come poke around.

24
GitHub Stars
5
Forks
MIT
License
7
AI Providers
HOSTED VERSION — COMING SOON

Want to skip the CLI?

We're building a hosted version of Clipify — upload from the browser, no local setup. Drop your email to get early access when it's ready.

No spam. Unsubscribe anytime. Your email is only used to notify you about Clipify.