Prompt With Jarvis
Work with meGet the free pack
All guides

VERTICALYou commented this keyword

SupoClip

An open-source AI tool that finds the best moments in a long video and renders them as 9:16 clips with captions, available hosted or self-hosted.

Before you start:
  • A web browser (hosted), or Docker plus API keys (self-host)
Setup
20 min
Level
Intermediate
Stars
1.3k
License
AGPL-3.0
Jump to the install commandView on GitHub

What it does

SupoClip takes a long video, like a podcast, talk or stream recording, finds the most clip-worthy moments, scores them, and renders them as vertical 9:16 clips. It adds face-centered cropping, word-synced subtitles and AI-written hook titles, with optional stock B-roll and transitions.

It is open-source and can run on your own hardware with no watermarks, or you can use the hosted web app and iOS app.

Use it (easiest)

  1. Go to supoclip.com and use the hosted web app; there is no infrastructure to run.
  2. Or install the iOS app from the App Store, which uses the same pipeline from your iPhone.
  3. Upload a long video and let it pick, score and render the clips.
  4. Use the built-in editor to trim, split and merge, then export with TikTok, Reels or Shorts presets.

Self-host it with Docker

  1. You need Docker, an AssemblyAI API key for transcription, and one LLM provider key (Google, OpenAI, Anthropic or a local Ollama model).
  2. Clone the repo and create a `.env` in the root with your keys (see the sample below).
  3. Run `docker-compose up -d`. First startup takes a few minutes; watch it with `docker-compose logs -f`.
  4. Open http://localhost:3107, create an account, and start clipping.

Clone and env file

Terminal
git clone https://github.com/FujiwaraChoki/supoclip.git
cd supoclip

# create .env in the root with:
# ASSEMBLY_AI_API_KEY=your_assemblyai_api_key
# LLM=google-gla:gemini-3-flash-preview
# GOOGLE_API_KEY=your_google_api_key

docker-compose up -d

Good to know

There is also an MCP server in the repo's mcp/ folder, so you can drive SupoClip from Claude, Cursor and other MCP clients (an MCP server is a connector that gives Claude a new tool).

Self-hosting needs paid API keys (AssemblyAI plus an LLM provider), though you can run fully local with Whisper and Ollama per the self-hosting tutorial. The project is licensed AGPL-3.0.

Links

Want the next one first?

One new guide a week, straight to your inbox.

More like this