Trending AI projects

A MATLAB and GNU Octave library for designing and testing marine control systems, featuring 6-DOF vessel models and comprehensive guidance, navigation, and control (GNC) functions.

Perch is an AI‑powered semantic linting CLI that scans code, asks decision‑model APIs to reason about potential defects, and returns ranked, confidence‑scored issues. It supports custom natural‑language rules, works with multiple model providers, and integrates with CI and assistant tools.

An open-source cloud architecture diagram generator that creates professional AWS, Azure, and GCP diagrams from either plain-English AI prompts or Terraform code.

Cortex turns API specifications and Markdown into typed SDKs, interactive documentation, and MCP servers for AI agents from a single project configuration.

A comprehensive cross-platform client for the Xiaozhi AI assistant, supporting real-time voice and text interaction across mobile, desktop, and AI hardware terminals.

InspireSkill is an agent-native CLI and toolkit that allows AI coding assistants to autonomously manage resources, jobs, and workflows on the Inspire compute platform.

Vidu S is a research system for real‑time, interactive video generation and live‑stream editing. Vidu S1 (540p) reacts to spoken commands; Vidu S2 (720p) adds high‑resolution avatars, reference‑image editing, and stereo output for VR. Speed comes from self‑replay forcing training and TurboDiffusion/TurboServe inference optimisations, enabling 25‑42 FPS on consumer GPUs. The project provides an online demo, REST API, and an agent‑skill for easy integration, with full English/Chinese docs and arXiv papers.

ChatArchiveGuard is a Python‑only CLI that locally scans chat export files (text, JSON/JSONL, SQLite) for secret/PII patterns, format errors, and SQLite integrity, reporting aggregate counts and coverage flags without modifying or uploading data.

A htop-like task monitor for GPUs and AI accelerators that provides real-time visual monitoring of utilization, temperature, and power across multiple hardware vendors.

trellis.cpp is a C++ GGML implementation of Microsoft’s TRELLIS 2‑4B image‑to‑3D pipeline. It converts a single image (or a Stable Diffusion‑generated image) into a fully textured GLB model, running entirely in native C++ with optional GPU back‑ends (CUDA, Vulkan, ROCm, Metal, Qualcomm HTP). It provides a CLI, a Tauri‑based desktop UI (Trellis Studio), and an HTTP server for programmatic use. The repo includes the full pipeline – background removal, three flow‑transformer stages, VAE decoders, mesh extraction, QEM decimation, UV unwrapping and WebP‑compressed PBR textures – and ships pre‑built binaries and GGUF model files.

A collection of reference samples for CUDA developers that demonstrate the features of the CUDA Toolkit across C++ and Python.

A self-hosted AI roleplay chat platform that allows users to run their own server for multi-device access and centralized data storage.

A visual prompt and LoRA assistance suite for ComfyUI designed for anime AI art, featuring searchable databases of 40k+ artists, characters, and clothing.

An automated video generation tool that creates scripts, synthesizes voiceovers with local voice cloning, and matches relevant video clips with synchronized word-level subtitles.

TurboVLA is a compact vision‑language‑action model for robot manipulation that directly fuses visual and textual inputs (via DINOv3 and BERT) and predicts continuous action chunks. It runs at 32 Hz on an RTX 4090 using < 1 GB VRAM, achieving 97.6 % success on LIBERO and 88 % on RoboTwin 2.0. The repo provides installation scripts, training/evaluation commands, and pretrained checkpoints.

A 3D virtual video store that turns digital media libraries from Jellyfin, Plex, or Emby into walkable, nostalgic aisles of VHS and DVD rentals.

WeSight is an Obsidian plugin that integrates local AI code assistants (Claude Code, Codex, OpenCode) for chat, inline editing, Git sync, and Chinese social‑media publishing, while keeping most data on‑device.

A curated directory of open-source AI agent skills for growth hacking, marketing, and revenue operations, organized into 14 topic categories covering everything from SEO and email to paid media and app-store optimization. Edited by enso.bot, it links exclusively to external open-source resources and accepts community contributions via pull request.

A series of open generative 3D world models from NVIDIA that create 3D and 4D scenes from a single image or video.

agterm is a minimal macOS terminal that groups shells into named workspaces and exposes a full control API (`agtermctl`) for scripts and AI agents to create sessions, type commands, split panes, run overlays, and read back output. It integrates optional Claude/Codex skills so agents can manage their own layout and report status, and it supports restoring live sessions after a restart. Distributed as a signed app for Apple‑silicon Macs and via Homebrew.

A lightweight, high-performance text chunking library for RAG pipelines that offers multiple splitting strategies and extensive integrations with vector databases and embedding providers.

A framework for building federated AI systems that enables decentralized model training across various ML frameworks and devices.

A Model-as-a-Service (MaaS) platform and library that provides a unified interface for discovering, inferring, and fine-tuning hundreds of state-of-the-art AI models across multiple domains.

Everywhere is a cross‑platform desktop AI assistant (Windows/macOS) built with .NET 10 and Avalonia. It captures on‑screen context via accessibility APIs and lets you query LLMs (OpenAI, Anthropic, Gemini, etc.) through a global hot‑key, returning markdown‑rich answers. Features include browser/file/terminal agents, rich UI, multi‑language UI, and support for local Ollama models. The project is open source under its LICENSE and provides installers, a portable zip, and full build instructions.

A high-performance LLM inference acceleration engine developed by Alibaba that optimizes throughput and latency for production-scale deployments.

Company Research Agent is a FastAPI + React app that, given a company name, automatically searches the web (via Tavily), scores sources, and uses any OpenRouter‑compatible LLM (or OpenAI/Gemini) to produce a structured, source‑backed briefing, streamed live to the UI and exportable as PDF. Optional MongoDB persistence and Docker deployment are supported.

LitGPT is a high-performance, abstraction-free toolkit for pretraining, finetuning, and deploying over 20 popular LLMs at enterprise scale.

An API service and browser extension suite that extracts watermark-free images and videos from Doubao, Dola, and Qianwen chat conversations.

A curated RSS feed collection (718 feeds, 16 OPML bundles) with strict quality checks, paired with a macOS reader app featuring AI summaries, translations, podcast playback, and FreshRSS sync.

Apache AGE is a PostgreSQL extension that adds multi-model graph database capabilities, allowing users to perform graph queries alongside standard relational SQL.

Lorca is a self-hosted, end-to-end encrypted bot orchestration platform that runs bots on your own machines and lets them collaborate in group chats. It includes desktop apps, a CLI, and mobile apps, supporting multiple AI providers and offering a private alternative to cloud bot services.

A curated collection of 358 GPT-6 Astra prompts for building 3D games, scenes, and interactive worlds in Blender, Three.js, Unreal Engine, Unity, and the browser, with entries available in 14 languages and 12 containing source code.

A curated, evidence-based catalogue of repeatable architectural patterns and workflows for building production-ready autonomous AI agents.

An educational resource from Google DeepMind that explains how to scale Large Language Models on TPU hardware while avoiding common communication bottlenecks.

A high-performance programming language with TypeScript-like syntax designed for AI agents, featuring a compiler that provides actionable JSON edits to fix coding errors.

An autonomous, self-improving AI agent delivered as a single Rust binary that runs locally in the terminal with zero telemetry and multi-provider LLM support.

An open-source, offline audio transcription app for self-hosters that provides private, local transcription, speaker diarization, and AI-powered chat with transcripts.

A professional Persian writing and RTL document toolkit that removes "AI-sounding" prose and fixes formatting errors in AI-generated Farsi text.

A toolkit for merging pre-trained language models in the weight space, allowing the combination of model strengths without additional training or increased inference costs.

A lightweight block diffusion model for speculative decoding that enables efficient, high-quality parallel drafting to speed up LLM generation.

Chubby Skills is a Python‑based toolkit that lets creators capture videos, podcasts, articles and local docs, turn them into searchable Markdown with full source metadata, and export citation‑backed material packs for AI agents. It works offline by default, offers optional transcription and enrichment, and provides a lightweight MCP server for agent integration.

A reference implementation of a year‑long context‑management system for personal AI assistants. It provides templates for AI persona and user preferences, a three‑level memory store, and periodic scripts that ingest daily observations and produce weekly reflections. The repo is a blueprint—not a plug‑and‑play tool—intended for users who want to build their own context‑aware AI by filling in personal data and optionally adding public skill plugins.

Claude Office Visualizer is an open‑source, real‑time pixel‑art dashboard that animates Claude Code’s agent activity. It includes a FastAPI backend, a Next.js + PixiJS frontend, and hooks/plugins that send Claude events to the visualizer. Features include multi‑floor navigation, a cross‑session Command Center, 12 whiteboard display modes, Git status panels, and optional AI‑generated names/summaries via a Claude OAuth token. Install with `make install-all` and run `make dev-tmux` to view the office at http://localhost:3000, or use the provided Docker setup.

A C++/ggml inference port of NVIDIA's LocateAnything-3B that enables fast, open-vocabulary object detection on CPU and GPU without a Python runtime.

A cross-platform AI chatting application that provides advanced tools for roleplay, including multiple API support, long-term memory systems, and immersive visual assets.

CyberVerse is an open-source framework for building real-time digital-human agents that can see, hear, and talk via WebRTC. It combines persona memory, RAG, tool calling, and optional avatar video (from a single photo) to create voice-centered AI characters, with full-duplex interaction and multi-agent task delegation.

AvatarAI is an open‑source, production‑grade system that turns any portrait photo and a short voice sample into a real‑time, lip‑synced talking avatar. It streams Whisper‑transcribed speech to a selectable LLM (Claude, GPT‑4o, or local Ollama/vLLM), feeds the generated text to a multilingual TTS (Chatterbox/XTTS), and renders video frames with MuseTalk V1.5. The stack includes Next.js front‑end, FastAPI back‑end, PostgreSQL, Redis, Celery, Prometheus, and Docker‑Compose, with optional one‑click AWS GPU deployment. Features include zero‑shot voice cloning, barge‑in interruption, multi‑LLM support, JWT auth, conversation memory summarisation, and optional face‑restoration via GFPGAN.

A unified memory stack for LLM agents that provides semantically lossless compression for text and multimodal data, featuring a self-evolving retrieval system to maximize recall while minimizing token usage.