Today's AI Frameworks & SDKs: Fastest-Growing Projects — June 28, 2026
Today's AI Frameworks & SDKs category on GitHub has seen a diverse range of projects gaining traction, with notable growth across several innovative tools designed to enhance machine learning and reinforcement learning capabilities. One standout project is CortexPrism/cortex, which offers an open-source agentic harness system that supports the development of intelligent applications.
CortexPrism/cortex provides an open-source agentic harness system aimed at developing sophisticated AI-driven applications. Its impressive growth score of 41.86 and a steady increase in stars to 215 reflect its growing importance as developers seek more robust tools for building intelligent systems.
leyten/shard is another project that has caught the attention of GitHub users, with its pipeline-parallel LLM inference across GPUs on separate machines. This tool's growth score of 39.96 and rising stars to 373 indicate a growing demand for efficient distributed computing solutions in AI applications.
gateGPT by fguzman82 is a project that translates full Transformer models into custom hardware, specifically RTL for Virtex-5 FPGA, achieving impressive token processing speeds at about 56k tokens per second. With its growth score of 34.47 and stars increasing to 578, gateGPT highlights the importance of optimizing AI workloads for specialized hardware.
UniRL from Tencent-Hunyuan is a framework designed for unified multimodal model reinforcement learning, aiming to bridge the gap between different modalities in machine learning tasks. Its growth score of 33.17 and rising stars to 724 suggest that this project is gaining significant traction among researchers and developers looking to advance their work in multimodal RL.
caezium/Burrow offers a native macOS GUI for managing disk space, analyzing usage patterns, and optimizing performance through AI-driven insights. With its growth score of 32.79 and stars climbing to 813, Burrow demonstrates the growing need for user-friendly tools that leverage AI to enhance system management.
machinefi/trio-retina is a model-agnostic state layer designed to standardize event streams from various detectors into a unified format for easy querying and analysis. Its growth score of 26.12 and increasing stars to 118 suggest its utility in edge computing environments where real-time data processing is critical.
OtterMind/Nubase stands out as an open-source platform designed to convert AI-written code into practical applications, offering a suite of backend services for modern development teams. With a growth score of 25.82 and stars reaching 451, Nubase indicates the growing interest in AI-native tools that streamline application development.
Simpl3x3/ESEILANE is a high-performance knowledge graph engine tailored for next-generation intelligent applications, including LLMs and GraphRAG systems. Its growth score of 25.20 and rising stars to 105 showcase its importance in the realm of advanced AI-driven data management solutions.
john-rocky/coreai-model-zoo offers a community-driven model zoo and knowledge base for Apple's Core AI framework, focusing on iOS/macOS devices with detailed conversion guides and Metal kernel support. Its growth score of 24.00 and stars increasing to 225 highlight the growing interest in optimizing machine learning models specifically for Apple hardware.
databufflabs/databuff rounds out Today's selection as an open-source AI-native APM tool designed to monitor application performance, providing a cost-effective alternative to commercial solutions like DataDog or NewRelic. With a growth score of 19.95 and stars reaching 159, databuff illustrates the demand for accessible monitoring tools in the AI ecosystem.
These projects collectively demonstrate the dynamic nature of the AI Frameworks & SDKs space, with developers continuously seeking out new ways to optimize performance, enhance capabilities, and streamline development processes across a variety of platforms and use cases.
CortexPrism/cortex provides an open-source agentic harness system aimed at developing sophisticated AI-driven applications. Its impressive growth score of 41.86 and a steady increase in stars to 215 reflect its growing importance as developers seek more robust tools for building intelligent systems.
leyten/shard is another project that has caught the attention of GitHub users, with its pipeline-parallel LLM inference across GPUs on separate machines. This tool's growth score of 39.96 and rising stars to 373 indicate a growing demand for efficient distributed computing solutions in AI applications.
gateGPT by fguzman82 is a project that translates full Transformer models into custom hardware, specifically RTL for Virtex-5 FPGA, achieving impressive token processing speeds at about 56k tokens per second. With its growth score of 34.47 and stars increasing to 578, gateGPT highlights the importance of optimizing AI workloads for specialized hardware.
UniRL from Tencent-Hunyuan is a framework designed for unified multimodal model reinforcement learning, aiming to bridge the gap between different modalities in machine learning tasks. Its growth score of 33.17 and rising stars to 724 suggest that this project is gaining significant traction among researchers and developers looking to advance their work in multimodal RL.
caezium/Burrow offers a native macOS GUI for managing disk space, analyzing usage patterns, and optimizing performance through AI-driven insights. With its growth score of 32.79 and stars climbing to 813, Burrow demonstrates the growing need for user-friendly tools that leverage AI to enhance system management.
machinefi/trio-retina is a model-agnostic state layer designed to standardize event streams from various detectors into a unified format for easy querying and analysis. Its growth score of 26.12 and increasing stars to 118 suggest its utility in edge computing environments where real-time data processing is critical.
OtterMind/Nubase stands out as an open-source platform designed to convert AI-written code into practical applications, offering a suite of backend services for modern development teams. With a growth score of 25.82 and stars reaching 451, Nubase indicates the growing interest in AI-native tools that streamline application development.
Simpl3x3/ESEILANE is a high-performance knowledge graph engine tailored for next-generation intelligent applications, including LLMs and GraphRAG systems. Its growth score of 25.20 and rising stars to 105 showcase its importance in the realm of advanced AI-driven data management solutions.
john-rocky/coreai-model-zoo offers a community-driven model zoo and knowledge base for Apple's Core AI framework, focusing on iOS/macOS devices with detailed conversion guides and Metal kernel support. Its growth score of 24.00 and stars increasing to 225 highlight the growing interest in optimizing machine learning models specifically for Apple hardware.
databufflabs/databuff rounds out Today's selection as an open-source AI-native APM tool designed to monitor application performance, providing a cost-effective alternative to commercial solutions like DataDog or NewRelic. With a growth score of 19.95 and stars reaching 159, databuff illustrates the demand for accessible monitoring tools in the AI ecosystem.
These projects collectively demonstrate the dynamic nature of the AI Frameworks & SDKs space, with developers continuously seeking out new ways to optimize performance, enhance capabilities, and streamline development processes across a variety of platforms and use cases.