Trending AI projects

A unified multimodal foundation model that integrates image understanding, text-to-image generation, and instruction-guided image editing with a focus on spatial intelligence.

A full‑stack, chapter‑by‑chapter code base that implements Stefan Jansen’s *Machine Learning for Trading* textbook. It covers data ingestion, feature engineering, a wide range of ML models, rigorous back‑testing, risk & cost management, and production deployment (Docker, live‑broker adapters, MLOps). Nine case studies span ETFs, crypto, equities, options, FX, futures, and more, making it a practical learning platform for quantitative finance and AI‑driven trading.

A family of multilingual, multimodal translation models that support text, image, and audio inputs, featuring advanced instruction following for terminology, formatting, and syllable-controlled dubbing.

A facial recognition tool for macOS that allows users to unlock their Mac using their face by automatically typing the stored password.

An open-source, real-time infrastructure monitoring platform that provides per-second metrics and ML-powered anomaly detection with zero configuration.

An AI-powered video translation and dubbing tool that automates speech recognition, subtitle translation, and voiceover generation in a single pipeline.

cc‑haha is a cross‑platform desktop workbench (Electron + React) that lets you chat with Claude, ChatGPT, Grok and many other LLMs, manage multiple coding sessions, review diffs, preview web pages, generate images, run multi‑agent workflows, and even let agents control other desktop apps (Computer Use) – all from a single app.

Crush is a terminal‑based AI coding assistant that lets you chat with any LLM, keep per‑project sessions, use language‑server data, and extend functionality via custom Model Context Protocol services—all configurable through a Bash‑style `crushrc` and available on all major OSes.

Visual Enhancer is a Windows desktop (and CLI) app that lets you apply NVIDIA DLSS 5 Neural Rendering, DLSS Super‑Resolution, RTX Video SR, frame‑generation, HDR conversion, sharpening, grain and colour‑grading to images and videos. It offers a drag‑and‑drop, batch‑processing pipeline with real‑time previews and JSON presets, targeting creators who need AI‑assisted upscaling and enhancement on RTX GPUs.

rederive is a coding-agent skill that teaches you a codebase by making you re-design it yourself through guided questions, never direct explanations. It maps the repo into 5–10 design problems, tracks your progress in local files, and keeps everything on your machine with no telemetry.

A comprehensive, open-source knowledge base that provides a systematic learning path for AI infrastructure, covering everything from GPU hardware and CUDA programming to distributed training and large model inference optimization.

Hulu‑Med is an open‑source, multimodal medical vision‑language model that handles text, 2‑D images, 3‑D scans, and videos. It offers several model sizes (4 B‑235 B), full training code, and achieves state‑of‑the‑art results on dozens of medical benchmarks. The repo provides easy installation, HuggingFace‑compatible loading, and optional vLLM serving for high‑throughput inference.

A curated, searchable gallery of Figure 1/teaser images from six top AI/ML conferences (2023–2026), paired with FigureForge, a browser tool that retrieves reference figures via CLIP+BM25 and uses LLMs to draft a publication-style Figure 1 for your own paper.

MemT-1.0 is a memory foundation model that integrates dynamic local memory and trainable hyper memory to persist and evolve information across interactions, without needing original context at inference. It reduces storage for long histories and supports an experience-guided agentic coding framework, with two model sizes (Nano and Flash).

An AI-powered toolkit for travel hacking that enables agents to search award flights, compare points vs cash, and manage loyalty balances across dozens of programs.

An open deep research agent that automates the process of gathering information from the web and local documents to produce detailed, factual, and cited research reports.

Zotero MCP is an integration tool that connects your Zotero research library to AI agents and chat interfaces, allowing them to search, read, and annotate your academic papers.

A streaming dual-brain memory system for real-time voice agents that provides low-latency, personality-aware, and factual long-term memory.

A toolkit featuring a prompt gallery, agent skills, and a CLI for GPT Image 2 and 2.5 to simplify image generation, editing, and prompt extraction from reference images.

Omnigraph is an open‑source graph database built on the Lance columnar format that runs on any S3‑compatible object store. It offers typed schemas, multimodal retrieval (graph + vector + full‑text), Git‑style branching, and Cedar‑based security policies, all managed via Terraform‑like `cluster` files. Designed for long‑horizon AI agents, it lets multiple agents share a versioned world model and retrieve context in a single query.

AstrBot plugin that downloads, organizes, and automatically collects memes via a WebUI, supporting multi-pack usage rules, semantic search, and cloud sync.

Khoj is an open-source personal AI assistant that serves as a second brain, allowing users to chat with their own documents and the internet using various LLMs.

A Windows-based plugin for DSH that automates the installation of custom prompt rules and manages necessary runtime dependencies with a simple one-click interface.

A blog series turned book that teaches how to build AI agents from first principles, centering on the idea that an LLM API is just a stateless JSON array and a 'harness' is the program that manages that array. It develops a toy agent in Python across 16 chapters, each addressing a concrete failure.

Flint is a Microsoft‑backed open‑source visualization language that lets AI agents (via MCP or a JavaScript library) turn simple, human‑editable specs into interactive charts across five rendering back‑ends, with built‑in themes and interaction presets.

OpenLore is a centralized, governed knowledge server that allows multiple AI agents to share, read, and update documentation using standard shell tools and the Model Context Protocol.

An AI-powered tool that helps users create tailored resumes and cover letters by matching a master resume against specific job descriptions using LLMs.

AdCraft is an open‑source, AI‑orchestrated platform that turns a textual brief into a fully editable advertising video. It chains specialized AI agents (creative director, scriptwriter, character designer, storyboarder, video director, sound director, etc.) to generate scripts, product renders, characters, scenes, storyboard images, video clips, music, and a final composited timeline. Users can interact via a visual canvas or natural‑language chat, edit any step, reuse assets from a unified library, and even clone existing ads to create new variants. The system supports many image/video/audio generation models and provides ready‑made character and scene catalogs.

CodexHost is a desktop extension that lets you run multiple AI coding agents (Pi, Claude Code, Grok, etc.) inside the official Codex Desktop IDE, offering unified diff view, cross‑agent delegation, remote execution, and live usage stats.

An open-source, event-driven orchestration platform for data, AI, and infrastructure workflows that unifies scheduled and real-time automation using a declarative YAML interface.

gentle‑shell is a terminal‑based coding‑agent built on the Pi LLM platform. It lets you describe a task in natural language, orchestrates the work through a structured “Organic Driven Development” workflow, can delegate to specialised sub‑agents, shows live Git diffs and agent status, and requires you to approve the exact change before committing. Installers exist for macOS, Windows and Linux, and the UI is customisable via a command palette and a `/gentle:customize` panel.

Ultimate Vocal Remover is a desktop application that uses advanced AI models to extract vocals from audio files, providing a simple interface for high-quality music source separation.

Arize Phoenix is an open‑source AI observability platform for LLM apps, offering tracing, evaluation, dataset/experiment tracking, prompt playground, and an AI‑assistant (PXI). It installs via pip/conda, runs locally or in Docker/Kubernetes, and integrates with major LLM frameworks and providers.

An Android application that enables local Stable Diffusion image generation with Snapdragon NPU acceleration for improved performance.

Vox Director is an AI‑driven agent skill that turns a one‑line topic into a short, Vox‑style collage explainer video (script, collage keyframes, animation, voice‑over, music, captions). It runs on Atlas Cloud models plus local ffmpeg, offering B‑roll, A‑roll, C‑roll, and presenter modes, and is installed as a skill for coding agents like Claude Code.

LlamaFactory is a unified, zero-code framework that enables users to fine-tune, evaluate, and deploy over 100 large language models through a simple Web UI or command-line interface.

RFD4-Proteina is a flow-matching foundation model for designing proteins and biomolecular complexes. It supports sequence–structure co-design, rich conditioning (motifs, hotspots, etc.), partial diffusion, and reward-guided search, with fast inference and multiple deployment options.

A versatile, open-source Android image editor that provides batch processing, background removal, and over 500 filter effects without tracking or proprietary dependencies.

A collection of standardized, installable skills that enable AI agents to perform automated SEO, content marketing, and copywriting tasks.

A family of open-source voice AI models for long-form speech recognition and synthesis, capable of handling up to 60-90 minutes of audio in a single pass.

h5i is a CLI workspace that lets AI agents red‑team web apps (browser automation, HTTP capture/replay, crawling) and also provides a Rust web framework (`h5i‑app`) whose code can be translated to Lean 4 for formal verification of security properties.

An open-source ChatGPT alternative that lets users run LLMs locally for privacy and control, or connect to cloud-based AI providers.

OpenGCUA is an open‑source stack that lets LLMs act as generalist computer‑use agents. It provides a macOS real‑device harness, a cross‑device Python/Rust SDK, and evaluation trajectories for eight frontier models, showing large gains in benchmark scores and efficiency.

The repository is a genuine software project that provides Docker images, deployment guides, and raw benchmark data for heterogeneous GPU inference of large mixture‑of‑experts LLMs (e.g., Qwen‑3.8‑Flash‑Next, DeepSeek‑V4.1‑Flash). It shows how to split prefill and decode across different GPU types (RTX 3080, RTX 6000D, DGX Spark, AI Max 395) and reports untuned performance numbers up to 16 k tok/s prefilling and ~442 tok/s decoding.

FreeToken is an edge‑native Mixture‑of‑Experts serving engine that lets you run frontier‑scale open‑weight models (e.g., 290 B‑parameter MoE) on a consumer PC. It combines CPU‑GPU co‑execution, double‑buffered streaming, semantic‑aware caching, and elastic VRAM management to deliver fast, low‑latency inference via a GUI or command‑line OpenAI‑compatible API.

DSH Mobile is a DeepSeek Harness plugin that adds a secure, real‑time mobile layer, letting you continue the same AI chat session from an Android phone (or browser) via LAN or various tunnel services. It provides touch‑friendly UI, device pairing, diagnostics, extension/customisation via `/mobile` commands, and strong security (private CA for LAN, trusted HTTPS for remote).

A collection of 54 open-source skills that teach AI coding agents professional motion graphics and animation workflows, featuring a deliver-and-verify loop for quality control.

An AI-powered command-line tool that automates the entire bug bounty lifecycle, from reconnaissance and vulnerability scanning to report generation.