An open-source AI content workbench for social media creators that automates the end-to-end workflow of trend discovery, planning, asset creation, and multi-platform publishing.

A real-time neural renderer that applies NVIDIA's DLSS 5 sharpening and enhancement effects to the entire Windows desktop or specific windows.

Lightpanda Browser is a lightweight, Zig‑written headless browser built for AI agents and large‑scale web automation. It runs a full V8 JavaScript engine, offers CDP and WebDriver BiDi interfaces, and includes an “agent” mode that lets LLMs control the browser via plain‑English tasks, recording deterministic JavaScript scripts. Benchmarks claim ~16× lower memory and ~9× faster page loads than headless Chrome. Installable via Homebrew, AUR, direct binaries, or Docker, it supports Linux/macOS (WSL2 for Windows) and integrates with major LLM providers. The project is MIT‑licensed, actively tested (unit, end‑to‑end, Web Platform Tests), and backed by a community Discord.

An open‑source Python tool that automatically creates vertical short clips from YouTube or local videos. It transcribes with local Whisper, uses an LLM (OpenAI, Gemini, or MuAPI) to rank viral moments, optionally adds an AI‑generated hook, and renders the clips via a simple CLI or a local web UI.

A series of compact, SOTA language models (1B-2B parameters) optimized for on-device deployment, featuring strong capabilities in coding, math, and agentic tasks.

An MCP server that enables AI agents to use thousands of Apify Actors for web scraping, data extraction, and automation.

A fly-connectome simulation that uses a biological neural network structure to trade cryptocurrency on Coinbase.

ai‑memory is a Rust‑based server that records every observation a coding‑assistant makes (prompts, tool calls, session boundaries) as plain markdown files in a git‑backed wiki. The server can be run locally, in Docker/Podman, or on a homelab box and is designed to be shared across many agents (Claude Code, Codex, Cursor, Gemini CLI, etc.), machines and team members. Hooks posted by agents are consolidated into readable wiki pages at session end; the next session receives a concise handoff so it can continue exactly where the previous one stopped. Storage is human‑readable markdown plus a SQLite index (FTS5, entity and optional vector search). By default it needs no LLM or embedding keys; optional Anthropic/OpenAI integrations add LLM‑generated summaries and semantic search. Multi‑user authentication, audit logs and per‑person attribution are built in. Install via Arch AUR, Docker wrapper, or native binary, start the server, register a client and its hooks, and the system works transparently.

A streaming dual-brain memory system for real-time voice agents that provides low-latency, emotionally aware long-term memory with minimal token consumption.

A semantic code intelligence platform that provides AI agents with a pre-built knowledge graph of a codebase to eliminate slow file-by-file discovery and reduce token costs.

AuK is a 1.5 B‑parameter open‑source foundation model for speech generation, editing, enhancement, and source separation. It works via natural‑language instructions, offers a high‑quality base model and a fast 4‑step distilled “Flash” version, and provides CLI, Gradio UI, Python API, fine‑tuning scripts, and ComfyUI integration.

An AI-powered file type detection tool that uses a deep learning model to accurately identify over 200 content types with high precision and speed.

An open-source rewriting pipeline that uses LLMs and multi-step cross-lingual translation to make AI-generated text feel more natural and bypass AI detectors.

An agent skill that uses a closed-loop cycle of image generation and AI critique to build visually impressive games, apps, or 3D scenes.

A universal provider proxy that lets you use any LLM (Claude, Gemini, Grok, etc.) with Codex, Claude Code, Claude Desktop, and Grok Build.

A macOS app that displays a screen-edge notch showing real-time usage limits and session status for various AI coding assistants.

anydoc is a pure‑Rust library (with Node, Python, and WebAssembly bindings) that converts 14 office‑document formats into consistent GitHub‑Flavored Markdown in a few milliseconds. It preserves document structure, supports optional hosted OCR for scanned PDFs, and is packaged as a CLI, an Agent Skill, and language‑specific SDKs, making it ideal for LLM‑oriented pipelines that need fast, uniform text extraction.

A plain C/C++ implementation for high-performance LLM and VLM inference across a wide range of hardware with minimal setup.

An end-to-end open source platform for machine learning that provides a flexible ecosystem of tools and libraries for building and deploying ML-powered applications.

Soup is a Python CLI (with optional web UI) that lets you fine‑tune LLMs—including 8B models—on low‑VRAM GPUs using a single YAML config and one command. It offers layer‑streaming, QLoRA, many training objectives, export to GGUF/ONNX/TensorRT, and an OpenAI‑compatible server, all with zero‑SSH, auto‑detected settings, and extensive documentation.

An open-source AI research agent that automates literature reviews, paper ranking, and scientific evidence synthesis across papers, web, and codebases.

An AI API gateway platform that allows users to distribute and manage AI subscription quotas by turning them into API-accessible services.

An edge-native MoE serving engine that enables running frontier-scale open-weight models on consumer hardware by optimizing CPU-GPU co-execution and memory management.

A portable C99 inference engine that enables a 2.78-trillion-parameter model to run on a single CPU with as little as 8 GB of RAM by streaming weights from disk.

An open-source, local-first AI voice studio for voice cloning, speech generation, and system-wide dictation, featuring an MCP server for AI agent integration.

A self-hosted, extensible AI platform that provides a unified interface for local and cloud-based models, featuring built-in RAG, agent creation, and enterprise-grade user management.

A CLI for git worktree management designed to let developers run multiple AI agents in parallel by simplifying the creation and switch between separate working directories.

TimesFM is a Google‑research foundation model for time‑series forecasting. The open‑source repo provides pretrained checkpoints (up to version 3.0), a Python library with PyTorch and Apple‑silicon (MLX) back‑ends, support for univariate and multivariate series with covariates, quantile forecasts, and LoRA‑style fine‑tuning. It achieves state‑of‑the‑art scores on major benchmarks, integrates with Google Cloud (BigQuery ML, Vertex Model Garden, Sheets), and is licensed under Apache‑2.0 for code but the latest weights are non‑commercial only.

OpenWhispr is a cross‑platform desktop app for privacy‑first voice dictation, meeting transcription, and AI‑assistant interaction. It supports offline speech‑to‑text with Whisper‑based models or cloud services, offers speaker diarisation, note‑taking with semantic search, team collaboration, and a public API for developers.

An open source local inference engine that profiles your hardware to recommend, download, and run the best-fitting LLMs locally and privately.

A WiFi sensing platform that turns radio signals into spatial intelligence for camera-free presence, vital signs, and pose estimation.

MyContext is a desktop app that continuously ingests a user’s work data (chat, docs, meetings, etc.) into a private, locally stored knowledge graph. The graph can be queried by LLMs, local agents, or authorized AI apps, providing them with a personal context layer for more informed responses. Features include a digital‑self assistant, natural‑language search with source citations, and strict human‑in‑the‑loop controls. The project is actively developed and released under the Elastic License 2.0.

An open-source library of curated prompts for AI chat models, featuring a web interface, self-hosting options, and prompt engineering educational resources.

Microduck is the control software for a tiny biped robot that uses reinforcement learning policies to walk, roll, and interact with its environment.

An AI-powered job-seeking workflow plugin for developers that transforms technical project logs and open-source contributions into professional resumes and interview preparation materials.

An AI agent skill that uses Remotion to turn product screenshots and copy into cinematic promo videos using a library of 157 shot recipe cards.

A lightweight, stealthy headless browser engine written in Rust for AI agents and web scraping, serving as a high-performance, low-memory replacement for headless Chrome.

A portable toolkit and agentic framework for automating the discovery, reproduction, and patching of software vulnerabilities to build more secure code.

A reference implementation of AI shopping and merchant agents built on Claude, providing frameworks for customer-facing commerce and secure back-office management.

A Codex Skill that transforms text or images into interactive 3D holographic cards with parallax effects, providing both a Three.js web viewer and an editable Blender project.

SkillSpector is an open‑source NVIDIA tool that scans AI‑agent skills (plugins, prompts, code) for security problems. It supports static analysis of 71 known vulnerability patterns, optional LLM‑based semantic checks, live CVE look‑ups, baseline suppression, and multiple output formats (terminal, JSON, Markdown, SARIF). It can run as a CLI, Docker container, or MCP server, making it useful for developers, CI pipelines, and production agents.

A specification framework that adds a lightweight requirement layer between humans and AI coding assistants to ensure alignment before code is written.

screenshot‑to‑code is an open‑source app that converts UI screenshots, mock‑ups, or short screen recordings into ready‑to‑run front‑end code (HTML/CSS, React, Vue, etc.) using LLMs such as Gemini, GPT‑5.5, and Claude, with optional asset extraction and visual preview.

A production-ready agent framework that provides an SDK and a FastAPI-based service to build, deploy and scale multi-agent applications with built-in RAG and tool management.

A browser designed for humans and AI agents to work in parallel, allowing agents to run automated tasks in isolated workspaces with access to the user's existing logins.

A real-time webcam application that detects facial expressions and hand gestures to trigger matching meme overlays, featuring a virtual camera for use in video calls.

A high-performance code intelligence engine that builds a structural knowledge graph of codebases to provide AI agents with efficient, low-token access to architectural insights.

A continual learning infrastructure for self-improving agents that automates the cycle of serving, collecting feedback, and updating model weights or agent harnesses.