Today's AI Frameworks & SDKs: Fastest-Growing Projects — April 18, 2026
This week, the AI Frameworks & SDKs space is abuzz with activity, as developers and researchers explore new ways to harness the power of artificial intelligence. We're seeing a surge in tools that enable decentralized AI computing, automated vulnerability research, and governed AI access, among other trends. As we dive into Today's top 10 fastest-growing repositories, it's clear that innovation is thriving in this space.
jxnxts/mcp-brasil takes the top spot with a growth score of 54.87 and 1,381 stars, offering an MCP server for 41 public Brazilian APIs. Its rapid growth can be attributed to its unique focus on providing access to formal verification for language models, making it a valuable resource for developers working with Brazilian APIs.
Agent-FM/agentfm-core comes in second with a growth score of 35.21 and 37 stars, boasting a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. Its growth is driven by the increasing demand for scalable and efficient AI computing solutions, as AgentFM allows users to run massive AI workloads directly across a global mesh of idle CPUs and GPUs.
yogthos/chiasmus takes third place with a growth score of 29.23 and 61 stars, offering an MCP server that gives language models access to formal verification. Its popularity stems from its ability to provide a robust and reliable framework for developers working on AI projects that require formal verification capabilities.
416rehman/DeepZero secures the fourth spot with a growth score of 24.73 and 91 stars, providing an automated vulnerability research framework that parses, decompiles, and analyzes thousands of Windows kernel drivers for exploitable IOCTLs using AI agents. Its growth is fueled by the need for more efficient and effective vulnerability research tools in the industry.
Yuan-lab-LLM/ClawManager rounds out the top five with a growth score of 24.52 and 576 stars, offering a Kubernetes-native control plane for AI agent instance management. Its popularity can be attributed to its ability to provide governed AI access, runtime orchestration, and reusable resources across multiple agent runtimes.
codefromkarl/ContextAtlas follows closely with a growth score of 20.93 and 25 stars, providing context infrastructure for AI coding agents through hybrid retrieval, project memory, and retrieval observability via CLI, MCP server, or embeddable library. Its growth is driven by the increasing demand for robust context management solutions in AI development.
OpenEnvision/Awesome-Multimodal-Modeling comes in seventh with a growth score of 20.36 and 248 stars, offering an extensive collection of multimodal modeling resources. Its popularity stems from its comprehensive coverage of MLLM, UMM, and NMM topics, making it a valuable resource for developers working on multimodal projects.
jnMetaCode/agency-orchestrator takes eighth place with a growth score of 19.05 and 323 stars, providing the first multi-agent framework that works with existing AI subscriptions without requiring an API key. Its growth is fueled by its ability to seamlessly integrate with popular AI services, making it an attractive solution for developers.
rhino-acoustic/NeuronFS secures the ninth spot with a growth score of 17.57 and 136 stars, offering an innovative file system that governs AI agents using B-tree NeuronFS. Its popularity can be attributed to its ability to provide efficient and scalable infrastructure for LLM agents.
Tencent-Hunyuan/HY-SOAR rounds out the top 10 with a growth score of 17.00 and 56 stars, providing self-correction for optimal alignment and refinement in diffusion models. Despite having only two commits in the past 30 days, its popularity stems from its innovative approach to improving diffusion model performance.
jxnxts/mcp-brasil takes the top spot with a growth score of 54.87 and 1,381 stars, offering an MCP server for 41 public Brazilian APIs. Its rapid growth can be attributed to its unique focus on providing access to formal verification for language models, making it a valuable resource for developers working with Brazilian APIs.
Agent-FM/agentfm-core comes in second with a growth score of 35.21 and 37 stars, boasting a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. Its growth is driven by the increasing demand for scalable and efficient AI computing solutions, as AgentFM allows users to run massive AI workloads directly across a global mesh of idle CPUs and GPUs.
yogthos/chiasmus takes third place with a growth score of 29.23 and 61 stars, offering an MCP server that gives language models access to formal verification. Its popularity stems from its ability to provide a robust and reliable framework for developers working on AI projects that require formal verification capabilities.
416rehman/DeepZero secures the fourth spot with a growth score of 24.73 and 91 stars, providing an automated vulnerability research framework that parses, decompiles, and analyzes thousands of Windows kernel drivers for exploitable IOCTLs using AI agents. Its growth is fueled by the need for more efficient and effective vulnerability research tools in the industry.
Yuan-lab-LLM/ClawManager rounds out the top five with a growth score of 24.52 and 576 stars, offering a Kubernetes-native control plane for AI agent instance management. Its popularity can be attributed to its ability to provide governed AI access, runtime orchestration, and reusable resources across multiple agent runtimes.
codefromkarl/ContextAtlas follows closely with a growth score of 20.93 and 25 stars, providing context infrastructure for AI coding agents through hybrid retrieval, project memory, and retrieval observability via CLI, MCP server, or embeddable library. Its growth is driven by the increasing demand for robust context management solutions in AI development.
OpenEnvision/Awesome-Multimodal-Modeling comes in seventh with a growth score of 20.36 and 248 stars, offering an extensive collection of multimodal modeling resources. Its popularity stems from its comprehensive coverage of MLLM, UMM, and NMM topics, making it a valuable resource for developers working on multimodal projects.
jnMetaCode/agency-orchestrator takes eighth place with a growth score of 19.05 and 323 stars, providing the first multi-agent framework that works with existing AI subscriptions without requiring an API key. Its growth is fueled by its ability to seamlessly integrate with popular AI services, making it an attractive solution for developers.
rhino-acoustic/NeuronFS secures the ninth spot with a growth score of 17.57 and 136 stars, offering an innovative file system that governs AI agents using B-tree NeuronFS. Its popularity can be attributed to its ability to provide efficient and scalable infrastructure for LLM agents.
Tencent-Hunyuan/HY-SOAR rounds out the top 10 with a growth score of 17.00 and 56 stars, providing self-correction for optimal alignment and refinement in diffusion models. Despite having only two commits in the past 30 days, its popularity stems from its innovative approach to improving diffusion model performance.