Empowering the future of multimedia

The Fluendo AI Engine is a high-performance, real-time video processing solution designed to eliminate the bottlenecks of traditional AI frameworks. By integrating the entire pipeline—from preprocessing to postprocessing—directly into a GPU-optimized, graph-based architecture, it achieves ultra-low latency across heterogeneous hardware. It is the ideal solution for developers needing decentralized, massively parallel execution on NVIDIA, AMD, or NPU-equipped systems.

Gst-WASM brings the industrial-grade power of GStreamer to the browser, enabling complex multimedia tasks that standard Web APIs cannot yet handle. This project provides a bridge for web developers to execute native-level video processing within Node.js or browser environments via WebAssembly. By porting these capabilities upstream, we are expanding the boundaries of what is possible in web-based media playback and manipulation.

Fluster is an essential Python-based testing framework designed to ensure decoder conformance across a wide array of global video standards, including H.265, AV1, and VVC. By comparing decoder implementations against proven test suites, it provides developers with a reliable CLI tool to validate performance and compliance. Originally focused on HEVC, it has evolved into a versatile tool that supports nearly every major video and audio format in the industry.

Gst-Audit modernizes the way developers interact with GStreamer by providing a real-time, visual interface for pipeline auditing and inspection. Built with Next.js and React Flow, it transforms complex introspection data into interactive topologies, allowing for live monitoring of running pipelines. This tool bridges the gap between low-level system data and high-level observability, making debugging and optimization significantly more intuitive.

Flumes simplifies media asset management by providing an automated system for scanning, detecting, and indexing media files within any directory. It monitors file system changes in real-time and stores deep technical metadata, such as GstCaps and stream information, into a searchable database. This ensures that large-scale media libraries remain organized and that technical specifications are always accessible for downstream processing.

Our Open-Source Plugins repository serves as a hub for specialized GStreamer components developed by Fluendo’s engineering team. Key highlights include Hype, a hybrid parallel encoder built for maximum throughput, and Fluttml, which provides precise parsing and rendering for TTML subtitle files. These tools reflect our “contribution-first” philosophy, providing the community with high-performance solutions for niche media challenges.

The Git Upstream Workflow is a specialized methodology and toolset designed to simplify the maintenance of complex open-source forks. It provides a structured process for managing pending contributions and bug fixes before they are officially merged into upstream repositories. This workflow is the backbone of our collaborative efforts, ensuring that our custom enhancements remain synchronized with the rapidly evolving open-source ecosystem.