Open video models are free to download. But actually running one takes more than that — structuring prompts correctly, managing the whole generation pipeline. I kept ending up between two bad options: pay per token on a closed platform, or spend a weekend wiring a model into something that actually works. So I built Fraime to do that part for me. It reads your hardware, picks an open model that actually fits it, structures the prompt, and runs the generation — through a plain API, or straight from your coding agents over MCP. Every model in the catalog is unconditionally free to use, no license traps. It's open source, and I'd like people to use it. Video is what it does now — image and audio generation are next. Happy to answer anything about the hardware-matching or the model catalog.

Fraime

Open source AI media generation platform

Open source AI media generation platform