We're building a flexible, open-source alternative to mainstream digital document signing platforms. Dive into our journey and the insights we've gathered on this blog.
At DocuSeal, we run a form field detection computer vision model on an NVIDIA GPU inside the Ruby process of our Rails monolith. No Python, no microservices, and no HTTP requests. GPU AI workloads fit in a single Rails codebase, resulting, for us, in a more reliable, faster, and easier to maintain AI inference pipeline, all done just with Ruby.
Running open-source large language models locally is easy. In this post, we'll explore how to set up and run LLMs on your machine using Ruby, and share practical code snippets for inference, prompting, and embeddings-without relying on external provider APIs.
Ever clicked a link and had the next page appear almost instantly? That magical, speedy experience is becoming easier to achieve thanks to the Chrome Speculation Rules API. This new browser feature helps make navigation feel instant with just a few lines of HTML.
Usually, CSS print styles are used when preparing a web page for printing, but they can also be valuable for generating PDF files. These CSS print styles optimize layout and formatting for printouts, enhancing the document's appearance and readability.