Welcome to the Pipeless documentation!
Pipeless is an open-source computer vision framework that ships all the features you need to create and deploy efficient computer vision applications that work in real-time. Pipeless is the easiest way to create a computer vision application.
Next.js (among others) allowed web developers to create applications quickly and easily. However, for computer vision there are no equivalent solutions. It is hard to get started and be productive in a short span of time.
Pipeless gives you all the required pieces in one. Simply write short functions that take frames and Pipeless will run them on events. You can easily load industry-standard models or bring your own and load it in one of the provided inference runtimes, they are run using CPUs or GPUs out-of-the-box.
We have taken technology and concepts from other fields, like serverless computing, to create what we think is the ideal solution for computer vision. We replaced the typical HTTP events that trigger serverless functions by frame events that trigger your processing code. This makes Pipeless to process streams in a highly parallelized environment and allows to efficently process multiple streams from several sources at the same time.
So in short, Pipeless makes your code to run extreamly fast without headaches, you just write simple functions.
Nobody likes to read docs, so you can start by running some of the examples and playing with their code. If you need more context on what is happening, you can come back to the getting started guide and learn from the basics.
In the rare case in which you are one of those persons who read the docs, we hope you enjoy the guides! Any feedback on the docs is really appreciated! You can also contribute to the docs by clicking on the links at the right menu.
Pipeless versioning follows the semver standard, except for versions annotated with
beta, which may include breaking changes without a major version bump.