Today's AI Frameworks & SDKs: Fastest-Growing Projects — May 10, 2026
Today's the AI Frameworks & SDKs space, we're seeing a surge in tools focused on serving large language models (LLMs) and enabling AI-assisted trading. Developers are also showing interest in frameworks that facilitate self-hosted, single-binary solutions for AI agents. Here's a closer look at the top 10 fastest-growing tools in this category.
1bananachicken/MaaNTE has taken the top spot with a growth score of 69.46 and over 1,039 stars. This tool is an MAA异环小助手 powered by the MAAF framework, offering users a QQ交流群 for discussion and support. Its rapid growth can be attributed to its unique functionality and strong community backing.
HKUDS-AI/polymarket-ai-trading boasts a growth score of 61.31 and 139 stars, despite relatively few commits in the past month. This AI-assisted paper trading tool for Polymarket offers OpenAI market insight, vector search, Kelly sizing, and more. Its popularity stems from its ability to provide valuable insights for traders looking to leverage AI.
noonghunna/club-3090 has garnered significant attention with a growth score of 59.50 and over 693 stars. This community-driven project provides recipes for serving LLMs on RTX 3090, supporting multiple engines like vLLM, llama.cpp, and SGLang. Its growth can be attributed to the increasing demand for efficient LLM deployment solutions.
warpdot-dev/craft-agents-oss has a growth score of 55.89 and 217 stars, with a notable number of commits in the past month. This open-source project offers an electron-based solution for crafting agents with multi-LLM support, thin-client architecture, and more. Its popularity stems from its versatility and ease of use.
alash3al/stash has secured a growth score of 41.97 and over 674 stars, driven by its innovative approach to persistent memory layers for AI agents. This self-hosted solution stores episodes, facts, and working context in Postgres, eliminating the need for cloud services. Its growth can be attributed to its unique value proposition.
oritera/Cairn boasts a growth score of 31.38 and over 802 stars, despite being a relatively new entrant in the space. This general-purpose state-space search engine has been validated on autonomous penetration testing, showcasing its potential for diverse applications. Its popularity stems from its novel approach to problem-solving.
TheTom/vllm-swift has garnered significant interest with a growth score of 26.56 and over 252 stars. This vLLM Metal plugin powered by mlx-swift enables high-performance LLM inference on Apple Silicon, making it an attractive solution for developers working with Apple devices. Its growth can be attributed to its optimized performance.
ZhiYi-R/moon-bridge has a growth score of 23.90 and over 84 stars, despite being a relatively new project. This moon-bridge tool serves as a conversion layer for Anthropic Messages formats, making it compatible with Codex's OpenAI Responses API. Its popularity stems from its ability to bridge the gap between different AI frameworks.
Mininglamp-AI/cider has secured a growth score of 21.77 and over 314 stars, driven by its innovative approach to W8A8/W4A8 inference on Apple Silicon. This MLX custom primitive unlocks unused INT8 TensorOps in M5, resulting in faster LLM prefill times. Its growth can be attributed to its optimized performance.
jegly/Box has a growth score of 21.04 and over 323 stars, despite being a relatively new entrant in the space. This private on-device AI suite for Android offers a range of features like llama.cpp, whisper.cpp, and NPU/TPU/GPU acceleration. Its popularity stems from its comprehensive set of tools and features.
These tools showcase the diverse range of solutions emerging in the AI Frameworks & SDKs space, from serving LLMs to enabling AI-assisted trading and self-hosted AI agents. As the landscape continues to evolve, we can expect to see even more innovative solutions addressing specific pain points and use cases.
1bananachicken/MaaNTE has taken the top spot with a growth score of 69.46 and over 1,039 stars. This tool is an MAA异环小助手 powered by the MAAF framework, offering users a QQ交流群 for discussion and support. Its rapid growth can be attributed to its unique functionality and strong community backing.
HKUDS-AI/polymarket-ai-trading boasts a growth score of 61.31 and 139 stars, despite relatively few commits in the past month. This AI-assisted paper trading tool for Polymarket offers OpenAI market insight, vector search, Kelly sizing, and more. Its popularity stems from its ability to provide valuable insights for traders looking to leverage AI.
noonghunna/club-3090 has garnered significant attention with a growth score of 59.50 and over 693 stars. This community-driven project provides recipes for serving LLMs on RTX 3090, supporting multiple engines like vLLM, llama.cpp, and SGLang. Its growth can be attributed to the increasing demand for efficient LLM deployment solutions.
warpdot-dev/craft-agents-oss has a growth score of 55.89 and 217 stars, with a notable number of commits in the past month. This open-source project offers an electron-based solution for crafting agents with multi-LLM support, thin-client architecture, and more. Its popularity stems from its versatility and ease of use.
alash3al/stash has secured a growth score of 41.97 and over 674 stars, driven by its innovative approach to persistent memory layers for AI agents. This self-hosted solution stores episodes, facts, and working context in Postgres, eliminating the need for cloud services. Its growth can be attributed to its unique value proposition.
oritera/Cairn boasts a growth score of 31.38 and over 802 stars, despite being a relatively new entrant in the space. This general-purpose state-space search engine has been validated on autonomous penetration testing, showcasing its potential for diverse applications. Its popularity stems from its novel approach to problem-solving.
TheTom/vllm-swift has garnered significant interest with a growth score of 26.56 and over 252 stars. This vLLM Metal plugin powered by mlx-swift enables high-performance LLM inference on Apple Silicon, making it an attractive solution for developers working with Apple devices. Its growth can be attributed to its optimized performance.
ZhiYi-R/moon-bridge has a growth score of 23.90 and over 84 stars, despite being a relatively new project. This moon-bridge tool serves as a conversion layer for Anthropic Messages formats, making it compatible with Codex's OpenAI Responses API. Its popularity stems from its ability to bridge the gap between different AI frameworks.
Mininglamp-AI/cider has secured a growth score of 21.77 and over 314 stars, driven by its innovative approach to W8A8/W4A8 inference on Apple Silicon. This MLX custom primitive unlocks unused INT8 TensorOps in M5, resulting in faster LLM prefill times. Its growth can be attributed to its optimized performance.
jegly/Box has a growth score of 21.04 and over 323 stars, despite being a relatively new entrant in the space. This private on-device AI suite for Android offers a range of features like llama.cpp, whisper.cpp, and NPU/TPU/GPU acceleration. Its popularity stems from its comprehensive set of tools and features.
These tools showcase the diverse range of solutions emerging in the AI Frameworks & SDKs space, from serving LLMs to enabling AI-assisted trading and self-hosted AI agents. As the landscape continues to evolve, we can expect to see even more innovative solutions addressing specific pain points and use cases.