AI & ML · Weekly GitHub chart, Wed, Nov 18
AI & ML repositories pooled from every language chart and re-ranked by stars gained. As of Wed, Nov 18, 6 AM GMT+9.
- 1
ZhuLinsen/daily_stock_analysis
LLM-powered multi-market stock analysis system
It collects market data, real-time news, and technical indicators across A-shares, Hong Kong, US, Japanese, Korean, and Taiwanese stocks and ETFs to generate AI decision reports. It supports cost-free scheduled runs via GitHub Actions, Docker, or local tasks with automated notifications.
+4,355in 7 days
- 2
JustVugg/colibri
Hierarchical inference engine for large MoE models
colibri is a C-based inference engine that runs large mixture-of-experts models on consumer hardware. It keeps dense layers and hot experts in RAM or on a GPU while streaming cold experts from disk.
+2,306in 7 days
- 3
CopilotKit/OpenBot
Self-hostable open-source workspace for AI coworkers
OpenBot is a modular, self-hostable workspace for running AI coworkers with your choice of models, agents, and tools. It runs as a standalone desktop app or on Kubernetes and lets users watch bots use browsers and tools while requiring approval before actions.
+2,035in 7 days
- 4
ggml-org/llama.cpp
C/C++ inference engine for LLMs and MLLMs
Built on top of the ggml library, llama.cpp enables LLM and MLLM inference with minimal setup across a wide range of hardware.
+1,706in 7 days
- 5
Anil-matcha/Open-Generative-AI
Open-source AI image and video generation studio
Open-source studio for generating AI images, videos, audio, and lip sync using models such as Flux, Midjourney, Kling, Sora, and Veo. It runs without content filters and supports self-hosting as well as desktop apps.
+1,436in 7 days
- 6
FareedKhan-dev/kimi-k3-in-c
Portable C99 inference engine for Kimi K3 on a single CPU
A portable C99 inference engine that runs the 2.78-trillion-parameter Kimi K3 model on a single CPU without BLAS, frameworks, or GPUs. It streams the 1.56 TB checkpoint from disk so it can run in 8.24 GB of RAM while producing byte-identical output across memory budgets.
+703in 7 days
- 7
opensandbox-group/OpenSandbox
Fast sandbox runtime for AI agents
OpenSandbox provides isolated environments for AI applications to execute code, run commands, manage files, and operate browsers or desktops. It lets you start locally with Docker and deploy to Kubernetes through a unified sandbox API.
+600in 7 days
- 8
FujiwaraChoki/supoclip
Open-source AI video clipping tool
SupoClip takes long-form videos such as podcasts or stream VODs and uses AI to extract viral-worthy moments into vertical 9:16 clips. It handles face-centered cropping, word-synced subtitles, hook titles, and optional B-roll.
+573in 7 days
- 9
ashhart/TensorFold
Zig-based LLM inference engine for Metal, CUDA, and Vulkan
TensorFold serves language models from a Zig binary on Apple Silicon and NVIDIA GPUs. It reads checkpoints, tokenizes requests, and runs Metal or CUDA kernels directly without requiring Python or MLX.
+570in 7 days
- 10
ggml-org/ggml
Lightweight C/C++ tensor library for machine learning
ggml is a simple, portable tensor library for machine learning with minimal setup. It is implemented in plain C/C++ without external dependencies.
+498in 7 days
- 11
orailnoor/cross-platform-llm-client
Cross-platform AI chat client for local and cloud models
PrivateLM is a cross-platform AI chat client built with Flutter. It supports offline on-device GGUF model inference on Android and iOS alongside unified access to cloud APIs.
+478in 7 days
- 12
a2ui-project/a2ui
Open-source generative UI framework for agents
An open-source generative UI framework where agents stream layout data and apps render it using native components in React, Angular, Lit, Flutter, SwiftUI, and Jetpack Compose. It expresses UI as schema-validated data rather than executable code to keep agent-generated experiences safe.
+473in 7 days
- 13
facebookresearch/faiss
Dense vector similarity search and clustering library
Faiss is a library for efficient similarity search and clustering of dense vectors, including datasets larger than RAM. It is written in C++ with Python/numpy wrappers and optional GPU implementations.
+470in 7 days
- 14
metadist/synaplan
Open-source multi-channel AI assistant platform
Synaplan is an open-source AI platform for building assistants on your own documents and deploying them across web, WhatsApp, email, Outlook, Nextcloud, and ownCloud. It lets you mix local and cloud models while displaying the cost of every answer.
+319in 7 days
- 15
kagent-dev/kagent
Kubernetes-native framework for building AI agents
kagent is a Kubernetes-native framework for building, deploying, and managing AI agents. It uses Kubernetes custom resources such as Agent, AgentTemplate, and ModelConfig to declaratively define agent behavior and runtime configuration.
+169in 7 days
- 16
opencv/opencv
Open source computer vision library
OpenCV is an open source computer vision library written in C++. It provides functionality for computer vision, image processing, and deep learning.
+141in 7 days
- 17
techjarves/Uncensored-Local-AI-Multiplatform
Cross-platform Flutter app for offline local AI models
A cross-platform Flutter application that runs open-source GGUF AI models directly on Android, iOS, and desktop devices without cloud dependency. All conversations stay on-device and work completely offline after downloading a model.
+125in 7 days
- 18
a2aproject/A2A
Open protocol for agent-to-agent communication
Agent2Agent (A2A) is an open protocol that enables AI agents built on different frameworks and running on separate servers to communicate and collaborate. It allows agents to work together on tasks without exposing their internal state, memory, or tools.
+108in 7 days
- 19
alibaba/MNN
Lightweight inference engine for on-device LLMs and Edge AI
MNN is a lightweight deep learning framework developed by Alibaba for on-device model inference and training. It also provides MNN-LLM and MNN-Diffusion to run large language models and Stable Diffusion locally.
+96in 7 days
- 20
leejet/stable-diffusion.cpp
Pure C/C++ diffusion model inference engine
A lightweight diffusion model inference engine written in pure C/C++ on top of ggml with no external dependencies. It supports image and video generation as well as image editing across models like Stable Diffusion, FLUX, Wan, and Qwen Image via CLI and an embedded Web UI.
+88in 7 days
- 21
agentscope-ai/agentscope-java
Java framework for distributed, long-running AI agents
AgentScope Java 2.0 is a production-ready Java framework for building distributed, long-running AI agents. It provides a unified event stream, tool-call permission gating, AOP-style middleware, and sandboxed execution.
+85in 7 days
- 22
ggml-org/Llama-macOS
macOS menu bar app for running local LLMs
Llama is a 4 MB native macOS menu bar app for running local LLMs on top of llama.cpp. It starts a local API server at localhost:9931/v1 and loads models on demand while unloading them when idle.
+76in 7 days
- 23
cogwheel0/conduit
Native mobile client for Open WebUI and self-hosted LLMs
Conduit is a Flutter-based native iOS and Android client for Open WebUI, OpenAI-compatible endpoints, Ollama, and Hermes agents. It renders chats, code blocks, Mermaid diagrams, and LaTeX natively while supporting reverse proxy authentication and on-device storage.
+41in 7 days
- 24
zml/zml
Zig and MLIR-based multi-hardware AI inference stack
ZML is a production AI inference stack built with Zig, MLIR, and Bazel. It compiles a single codebase directly to accelerators including NVIDIA, AMD, Intel, TPU, and Trainium without rewriting.
+22in 7 days