Trending AI projects

ResumeForge is an open‑source desktop app (React + FastAPI) that lets you store job postings, manage a fact‑based resume database, and use LLMs (or local models) to generate, edit, and export tailored resumes—all data stays on your computer.

A research codebase that implements Context Language Models—LLMs that treat their context as a mutable file, allowing dynamic read/write updates. Built on the Harbor framework, it includes zero‑shot, in‑context learning, and RL pipelines, plus tools for efficient serving and integration with benchmarks like ContextBench.

A set of extensions for the Model Context Protocol (MCP) that enables developers to build ChatGPT plugins with native-feeling UI components like sidebar entrypoints, custom file viewers, and interactive forms.

OpenDLSS‑NR is a Windows‑only Vulkan re‑implementation of NVIDIA’s DLSS 5 Neural Rendering network. It reproduces the exact 71‑block Swin‑ViT model (FP8 on tensor cores) and provides both a high‑performance PTX kernel path and a reference GLSL path, plus a Filament demo and a WebGPU browser version. Users supply their own model weights; the tool can benchmark, verify bit‑exactness against NVIDIA fixtures, and run interactive demos.

ClearAI is a DSH plugin that adds an epistemic loop to help LLMs build a verified domain ontology during research sessions. Install via `dsh plugin add clearai-dsh@0.5.6`, then select the ClearAI preset. The UI splits into Explore (process) and Ontology (graph) panes, automatically tracking hypotheses, predictions, tests, evidence, and conflicts. It’s a real, Apache‑2.0‑licensed project for trustworthy knowledge construction.

A Python package for statistical computation that provides tools for estimation and inference for a wide range of statistical models.

A Python library for financial econometrics that provides tools for modeling volatility, testing for unit roots and cointegration, and performing statistical bootstrapping.

A community-maintained Linux distribution of the Claude Desktop app that adds native package support, custom themes, multiple profile management, and enhanced features like IDE-style file navigation and third-party model integration.

XActions is a comprehensive, free, open-source toolkit for automating X/Twitter via CLI, browser scripts, and an MCP server that lets AI agents control X without an API key. It scrapes data, manages accounts, and supports AI-driven workflows with human approval.

md2wechat is a Go‑based CLI that converts Markdown into WeChat public‑account HTML, offers advanced layout modules, image generation, multi‑account handling, and a JSON‑driven discovery interface for AI agents. It works in a free “prompt‑only” mode or a paid Professional API mode that returns final HTML.

An open-source AI penetration testing and bug-bounty tool that uses machine oracles to independently verify and prove every vulnerability finding.

PM-Skills is a library of 68 professional product management frameworks and templates that allows you to trigger standardized, high-quality AI outputs via simple commands instead of manual prompt engineering.

A JavaScript library that enables running pretrained machine learning models directly in the browser using ONNX Runtime, eliminating the need for a server.

A high-performance RLHF framework that combines Ray and vLLM to scale reinforcement learning from human feedback for LLMs and VLMs.

A fast reinforcement learning library that uses custom research-backed algorithms to train high-performance, tiny models in seconds.

`lms` is a CLI that controls LM Studio (and compatible runtimes) to start/stop the local API server, list, load, and unload locally stored LLMs, and stream logs, enabling scriptable management of local AI models.

The Model Context Protocol (MCP) is an open standard that enables AI applications to connect to data sources and tools consistently. This repository provides the core infrastructure and a collection of official Microsoft servers to integrate services like Azure, GitHub, and Microsoft 365 with AI agents.

A browser automation tool that lets AI agents control your existing, logged-in Chrome browser via an extension, enabling seamless interaction with authenticated websites.

An agent-native operating system for robotics that allows developers to build and control hardware like quadrupeds and drones using Python and LLM-based commands.

Data‑Analysis‑Agent is a Flask‑based AI agent that turns natural‑language questions into SQL queries, runs them on uploaded files or connected databases, auto‑generates Plotly visualisations, and returns business insights. It works with any OpenAI‑compatible LLM, supports many chart types, and offers Windows installers, cross‑platform zip, or script‑based installation.

An enterprise-grade AI data analyst based on Spring AI Alibaba that converts natural language to SQL, performs deep Python analysis, and generates visual reports.

A centralized framework for materials science and quantum chemistry that provides high-speed, machine-learned interatomic potentials for atomistic simulations.

The GitHub Copilot app is a desktop client that lets you run Copilot‑powered AI agents in isolated git worktrees or cloud environments, manage issues/PRs across repos, and automate prompts. It provides visual “canvases” for agent output, parallel sessions, and an “Agent Merge” flow to auto‑merge PRs after checks. Installable on Windows, macOS, and Linux; requires a Copilot plan or your own model key.

An open‑source Python framework for training, evaluating, and running inference with many state‑of‑the‑art music source‑separation models. It provides config‑driven training (single‑GPU or multi‑GPU), LoRA fine‑tuning, a CLI, a small GUI, and scripts for ensembling and converting models to ONNX/TensorRT.

Audio Separator is a Python library/CLI that uses deep‑learning models (MDX‑Net, VR, Demucs, MDXC) to split music files into stems such as vocals, instrumental, drums, etc. It supports CPU, CUDA, Apple Silicon (MPS/CoreML), and experimental DirectML, offers Docker images, and provides a model catalog with filtering options.

JJYB AI VideoAutoCut is a Windows‑only desktop app that combines AI‑driven shot detection, script generation, multi‑engine TTS/voice‑cloning, timeline syncing and export (including CapCut drafts). It uses Flask + PyWebView for the UI, FFmpeg/OpenCV for media processing, and configurable cloud or local AI models. The tool is free for personal, non‑commercial use only.

OpenCode Quota is an npm‑distributed plugin for the OpenCode AI coding assistant that displays real‑time token usage, quota limits, and cost information for many LLM providers. It offers in‑app UI panels, slash commands, and CLI tools, letting users monitor and manage their AI token budgets across personal and enterprise accounts.

Deep Eye is an AI-powered penetration testing framework that automates vulnerability scanning, payload generation, and security report writing by orchestrating multiple LLM providers.

OmniDocBench is a comprehensive benchmark dataset and evaluation framework designed to assess the performance of document parsing models across diverse, real-world document types and complex layouts.

A genuine open‑source project that provides a library of AI video‑generation prompts, a 5‑stage prompt methodology, a web‑based prompt builder, and a Claude Code skill for turning ideas into ready‑to‑run multi‑shot prompts for models like Sora, Seedance, Veo, Kling, etc.

comfy‑cli is a Python CLI that automates installing, managing, and running the ComfyUI generative‑media engine. It handles custom node installation, fast dependency resolution, local or cloud workflow execution, and provides JSON‑friendly output for scripting or LLM agents.

Trackio is a free, local‑first experiment‑tracking library compatible with the WandB API. It stores logs in SQLite (optionally exported to Parquet), offers a CLI/SQL query interface for LLM‑driven analysis, and provides a Gradio‑style dashboard that can run locally or be published to a Hugging Face Space. Features include alerts, webhooks, high‑throughput non‑blocking logging, and easy embedding of dashboards.

A family of parametric statistical 3D human models, starting with GNM Head, for high-fidelity representation of human geometry and appearance.

Super Agent Party is a cross‑platform AI desktop companion that lets LLM agents appear as 3‑D or Live2D avatars, control the computer, browse the web, and act as bots on many messaging and streaming platforms. It ships as ready‑to‑run binaries for Windows, macOS (M‑chip), Linux AppImage, Docker images, and source code, and offers an OpenAI‑compatible API and extensible plugin system.

A tool that converts Bilibili videos and posts into structured Markdown notes by archiving subtitles, comments, and keyframes for use in personal knowledge bases.

An MCP server that integrates Microsoft Teams with LLMs, allowing AI agents to read messages, create threads, and mention members.

A collection of reusable plugins, skills, and agent configurations that integrate security tools and workflows into AI agent platforms like Codex and Claude Code.

An automated vulnerability discovery system that uses a multi-agent AI framework to screen projects, audit source code, and generate CVE reports.

A Kubernetes-native distributed AI platform for scalable LLM fine-tuning and training across multiple frameworks like PyTorch, JAX, and DeepSpeed.

A collection of guides, tutorials, and code examples designed to help developers build, fine-tune, and deploy Google's open Gemma AI models.

A compiler and runtime system that reduces LLM inference latency by fusing multi-GPU computation and communication into a single high-performance megakernel.

Grounded Docs MCP Server is an open‑source, locally‑run service that scrapes and indexes documentation (websites, GitHub, npm, PyPI, local files) for any library version, stores it (optionally with vector embeddings), and exposes a MCP‑compatible HTTP endpoint. AI coding assistants (Claude, Copilot, Gemini, etc.) can query the server to get exact, up‑to‑date answers, reducing hallucinations while keeping data private.

A curated collection of 26 end-to-end machine learning projects spanning healthcare, computer vision, and NLP, designed to help developers build and deploy practical AI applications.

AgentSight is a kernel‑level observability tool for AI agents. Using eBPF and TLS tracing, it records prompts, model calls, subprocesses, file and network activity without requiring any SDK or proxy. It provides live dashboards (`top`), replay visualisations (`vis`), queryable SQLite logs, and OpenTelemetry export, helping developers debug failures, audit security, and understand token/resource costs of agents such as Claude Code, Gemini CLI, and Python‑based tools.

CubeStudio is an open‑source, MIT‑licensed, cloud‑native AI platform that covers the full MLOps lifecycle—from data ingestion and multimodal annotation, through notebook‑based development and drag‑and‑drop pipelines, to distributed training (including large‑model fine‑tuning) and zero‑code inference serving. It runs on Kubernetes, supports many Chinese‑made accelerators (GPU, NPU, vGPU), offers multi‑tenant resource scheduling, a compute marketplace, built‑in billing, and extensive monitoring.

Symfony AI is a collection of PHP components and bundles that allow developers to integrate AI platforms, agents, and data storage into their Symfony applications.

A library of evaluations built using the Inspect AI framework to measure the capabilities and risks of frontier AI models.

Accelerate is a thin PyTorch wrapper that removes the boilerplate for running training scripts on multiple GPUs, TPUs, or with mixed precision. You keep your own training loop and add just a few lines (an Accelerator object, a prepare() call, and accelerator.backward()). It also offers an optional CLI (accelerate config / accelerate launch) and a notebook launcher. It supports CPU, single/multi-GPU, multi-CPU, TPU, FP16/BF16/FP8 mixed precision, plus experimental DeepSpeed, FSDP, and Megatron-LM. Use it if you want control over your training loop without distributed-setup hassle; skip it if you want a fully high-level framework. Requires Python 3.11+ and PyTorch 2.6.0+; install via pip install accelerate.