In short
After five months of development, the major release of FFmpeg didn't make much of a splash, but this package is actually what powers most media pipelines in the industry. What engineers need to know.
FFmpeg 9.0 was released after five months of development—and the news went almost unnoticed. Meanwhile, this is a major release of the package that quietly handles the lion’s share of media processing in production: recording, transcoding, and decoding video and audio.
For a blog about development and AI, this is an important context. When you’re building pipelines with video and audio inference, collecting datasets from media, or routing streams through agents, FFmpeg usually ends up somewhere at the bottom of the stack—not as a trendy dependency, but as infrastructure without which nothing will work.
The release is written in C and distributed under the LGPL and GPL. The details of the changes in the source code haven’t been disclosed, but the fact that it’s a major release means that if you have a pinned version in your CI or Docker image, you should check for compatibility before someone updates the base layer.
The main takeaway: tools that actually carry the load are often updated without press releases or hype. FFmpeg is exactly the kind of case where a quiet major release deserves an engineer’s attention more than a high-profile presentation of yet another SaaS product.