Today's AI Frameworks & SDKs: Fastest-Growing Projects — May 08, 2026
Today's AI Frameworks & SDKs, we're seeing a surge in tools focused on serving large language models (LLMs) efficiently and effectively. The top-growing repositories are tackling challenges such as inference acceleration, state-space search, and AI-assisted trading. These innovative solutions are gaining traction among developers, with several projects boasting impressive growth scores and star counts.
1bananachicken's MaaNTE takes the top spot with a growth score of 75.04 and 931 stars. This tool is an assistant powered by the MAAF framework, offering features that have drawn in a large community. Its high commit activity over the past month suggests active development and user engagement.
warpdot-dev's craft-agents-oss follows closely with a growth score of 73.57 and 229 stars. This project provides an electron-based SDK for crafting AI agents, supporting multiple LLMs and offering features like OAuth and desktop AI integrations. Its growing popularity may be attributed to its versatility and ease of use.
HKUDS-AI's polymarket-ai-trading boasts a growth score of 72.36 and 140 stars. This tool enables AI-assisted paper trading on Polymarket, leveraging OpenAI market insights and vector search capabilities. Its growth suggests interest in using AI for trading and financial analysis.
noonghunna's club-3090 has a growth score of 66.95 and an impressive 622 stars. This community-driven project provides recipes for serving LLMs on RTX 3090 GPUs, supporting multiple engines and models. The high commit activity and star count indicate strong interest in optimizing LLM performance.
alash3al's stash rounds out the top five with a growth score of 47.68 and 666 stars. This tool offers a persistent memory layer for AI agents, storing episodes, facts, and working context in Postgres. Its growth may be driven by its ability to enhance agent performance and efficiency.
Other notable projects include oritera's Cairn, a state-space search engine validated on autonomous penetration testing; TheTom's vllm-swift, a high-performance LLM inference plugin for Apple Silicon; and jegly's Box, a private on-device AI suite for Android. These tools demonstrate the diversity of innovation in the AI Frameworks & SDKs space.
ZhiYi-R's moon-bridge has a growth score of 26.27 and 59 stars. This project serves as a conversion layer between Anthropic Messages formats and OpenAI Responses API, enabling compatibility between different AI systems. Its growth suggests interest in integrating multiple AI platforms.
Mininglamp-AI's cider boasts a growth score of 21.32 and 241 stars. This tool unlocks unused INT8 TensorOps on Apple Silicon for faster LLM prefill, built as MLX custom primitives. Its growth may be driven by the increasing demand for optimized AI performance on specific hardware.
1bananachicken's MaaNTE takes the top spot with a growth score of 75.04 and 931 stars. This tool is an assistant powered by the MAAF framework, offering features that have drawn in a large community. Its high commit activity over the past month suggests active development and user engagement.
warpdot-dev's craft-agents-oss follows closely with a growth score of 73.57 and 229 stars. This project provides an electron-based SDK for crafting AI agents, supporting multiple LLMs and offering features like OAuth and desktop AI integrations. Its growing popularity may be attributed to its versatility and ease of use.
HKUDS-AI's polymarket-ai-trading boasts a growth score of 72.36 and 140 stars. This tool enables AI-assisted paper trading on Polymarket, leveraging OpenAI market insights and vector search capabilities. Its growth suggests interest in using AI for trading and financial analysis.
noonghunna's club-3090 has a growth score of 66.95 and an impressive 622 stars. This community-driven project provides recipes for serving LLMs on RTX 3090 GPUs, supporting multiple engines and models. The high commit activity and star count indicate strong interest in optimizing LLM performance.
alash3al's stash rounds out the top five with a growth score of 47.68 and 666 stars. This tool offers a persistent memory layer for AI agents, storing episodes, facts, and working context in Postgres. Its growth may be driven by its ability to enhance agent performance and efficiency.
Other notable projects include oritera's Cairn, a state-space search engine validated on autonomous penetration testing; TheTom's vllm-swift, a high-performance LLM inference plugin for Apple Silicon; and jegly's Box, a private on-device AI suite for Android. These tools demonstrate the diversity of innovation in the AI Frameworks & SDKs space.
ZhiYi-R's moon-bridge has a growth score of 26.27 and 59 stars. This project serves as a conversion layer between Anthropic Messages formats and OpenAI Responses API, enabling compatibility between different AI systems. Its growth suggests interest in integrating multiple AI platforms.
Mininglamp-AI's cider boasts a growth score of 21.32 and 241 stars. This tool unlocks unused INT8 TensorOps on Apple Silicon for faster LLM prefill, built as MLX custom primitives. Its growth may be driven by the increasing demand for optimized AI performance on specific hardware.