Open-source video clipper with full control of the workflow

SupoClip brings source import, transcription, candidate selection, captions, framing, and exports into a browser workflow. The application is available under AGPL-3.0. Self-hosting is useful when you want to adapt the pipeline, choose providers, automate jobs with the REST API or bundled MCP server, or keep the application alongside your own systems. You still review each suggested cut before using it: access to the source does not make an automatic selection editorially correct.

The documented Docker Compose stack includes the frontend, backend, worker, PostgreSQL, and Redis. Begin with the self-hosting guide and a short video you know well. Verify that a task moves through processing, that a finished clip plays, and that captions and framing match the source. The examples above show the kind of output to inspect; they are not a benchmark for your hardware, models, or footage. Do a small first run before processing a backlog.

Running the app on your server does not automatically keep every part of processing on that server. Hosted transcription or language-model providers receive the inputs required for their work. For local speech transcription, configure Whisper; for local clip selection, configure Ollama and a suitable installed model. Check connectivity from the worker as well as from your terminal. Review your selected configuration before processing material that needs to remain on your infrastructure.

The code has no subscription fee for self-hosting, but hardware, electricity, persistent media storage, backups, maintenance, and optional API usage still have costs. Keep secrets private, review updates, monitor workers, and retain backups before upgrades. License terms also matter when you modify or distribute the application or provide a network service. Consult the repository license for the exact requirements. Choose managed hosting if operating this stack would distract from making clips.

If you need an interface for reviewing clips, timing captions, and correcting crops, evaluate the complete SupoClip app. If you are building a custom processing service, a library may be a better starting point. Automatic silence removal is another distinct editing job; it does not by itself decide which story or explanation makes a good social clip. Compare tools against the task you need to finish instead of assuming every open-source video project replaces the same workflow.

Test a recording you have permission to use and know well. Check complete thoughts, speaker changes, proper nouns, visual context, and final exports. Record setup effort separately from processing time, and include human review and provider costs in your decision. You can use paid hosted SupoClip to focus on creating clips or self-host to control the application and its configuration. Both routes benefit from checking the real output before adopting a repeatable workflow.