Trending AI projects

Irodori‑TTS is a flow‑matching Japanese text‑to‑speech model that supports zero‑shot voice cloning, caption‑based style control, LoRA fine‑tuning, speaker‑inversion embeddings, a distilled “MeanFlow” fast‑generation variant, and quantized checkpoints. It ships CLI tools, Gradio UIs, and training scripts for full‑model or adapter training.

Memorix is a Node‑JS CLI/library that adds a local, SQLite‑backed, searchable memory store to any Git project. It lets AI coding assistants (Claude Code, Copilot, Cursor, etc.) share project facts, code snapshots, design rationales and commit‑derived knowledge across sessions and tools via the Model‑Context‑Protocol. Features include autopilot worksets, observation/reasoning memory, Git‑memory ingestion, code‑graph indexing, orchestration of multi‑agent tasks, a web dashboard, and optional sync. Install with npm, run `memorix setup --agent <agent>` and the agents can start reusing the shared memory.

Paca is an open‑source, self‑hosted project‑management platform that integrates AI agents as first‑class Scrum teammates. It provides a shared Scrumban board, in‑app AI chat, BDD collaboration, role‑based IAM, WASM sandboxed plugins, and an MCP server that lets any compatible AI agent read/write workspace data. Installable via a single Docker‑Compose script or a Helm chart, it’s designed for teams that want AI deeply embedded in their workflow rather than added as a peripheral bot.

OpenWAM is an open, modular research stack for developing World-Action Models (WAMs) that combines video understanding with action generation. It provides infrastructure for controlled experiments, a set of derived design principles, and a large pretrained model (OpenWAM-alpha) trained on 518.5M frames of egocentric human and robot data, enabling researchers to systematically compare and improve their own models.

A tool that converts local WeChat chat histories into a searchable personal intelligence hub to track business opportunities, commitments, and key discussions.

A complete Obsidian vault template that transforms the app into a Life OS, integrating structured productivity workflows with an AI assistant and MCP bridge for automated life management.

openOMSI is an open‑source, Rust‑based re‑implementation of the OMSI 2 bus simulator that runs on modern graphics APIs, supports Windows/macOS/Linux/Android, and works with existing OMSI 2 maps, vehicles and mods.

An open-source observability platform that stores telemetry data in S3 and uses AI agents and natural language queries to simplify log and metric analysis.

A tool that parses multi-language codebases into a knowledge graph to enable natural language querying, surgical code editing, and dead code detection.

A library for mechanistic interpretability that allows researchers to reverse-engineer generative language models by inspecting and manipulating their internal activations.

The Agentics is a family of reusable GitHub Agentic Workflows that automate common repository tasks like issue triage, code review, documentation updates, and more, saving maintainers time and keeping projects healthy.

A modular, expert-level code review framework for Claude Code that provides structured, consistent feedback across 20+ programming languages and frameworks.

A desktop application that wraps the Pi AI agent into a native, installable interface for macOS, Windows, and Linux, simplifying session management, task scheduling, and local file interaction.

A collection of measurement-first agent skills that enable AI coding agents to audit and optimize web projects for performance, accessibility, and SEO using tools like Google Lighthouse.

An AI-powered QA engineer that automates testing for web and mobile apps by executing plain-language user journeys and maintaining a persistent, self-improving memory of application behavior.

Gondolin is a Node SDK and CLI that runs AI‑generated code inside fast, disposable Linux micro‑VMs (QEMU or experimental krun). It lets you enforce programmable network and filesystem policies, inject secrets only for allowed destinations, snapshot and resume VMs, and expose guest services via an ingress gateway. Ideal for safely executing agent‑produced scripts, web‑scraping, or sandboxed browser automation.

Xiaobei is an open‑source AI agent (built on openclaw) that automates content creation, video production, multi‑platform publishing, analytics, and 24/7 customer support for small‑to‑medium businesses—all controlled via a WeChat chat interface.

A Python library for audio and music signal processing that provides the foundational tools for building music information retrieval systems.

A centralized management tool for AI agent skills that allows users to install once and sync them across dozens of different AI coding tools and devices.

The MATLAB Agentic Toolkit is a MathWorks add‑on that connects LLM coding assistants (Claude Code, Copilot, Codex, Gemini, etc.) to a running MATLAB session via a local MCP server. It installs the server, loads curated “skills” covering core MATLAB and dozens of toolboxes, and exposes tools (run code, run tests, static analysis, list toolboxes) that the AI can call. This reduces hallucinations and token use, letting agents generate, test, and debug MATLAB code reliably.

This is a repository of ready-made AI subagent definitions for OpenAI's Codex, covering a wide range of development tasks with smart model selection and sandbox controls.

Unofficial Perplexity AI client with account automation, MCP server, sync/async APIs, and browser-based account generation to bypass query limits.

A command-line automation tool for Xiaohongshu that enables programmatic content publishing, account management, and data retrieval via the Chrome DevTools Protocol.

MCP for Adobe Premiere Pro is a local Model Context Protocol server that connects AI assistants like Claude and Cursor to Adobe Premiere Pro, enabling AI-driven editing, project organization, and export verification through Adobe's documented APIs.

Backburner is a distributed inference framework that allows users to run large LLMs by offloading model layers and context memory from a Mac to one or more iPhones or iPads over USB-C.

A bridge extension for the pi coding agent that allows users to run Claude models and delegate tasks to Claude Code while maintaining a unified terminal workflow.

OpenAdapt is an open‑source Python engine that records a human’s UI interaction (desktop, web, RDP, Citrix), compiles it into a repeatable program, and then **verifies** that the data entered into the target system was saved correctly by reading the record back through an independent channel. If verification fails, the run halts and a human is prompted. It supports multiple surfaces, policy‑driven certification, privacy‑first data handling, and can be run locally or via a managed cloud service.

Papr is a native desktop RSS reader plus a command‑line tool built for AI agents. The desktop app offers offline‑first reading, tagging, full‑text extraction, audio playback, and optional LLM‑generated summaries or Q&A. The `papr` CLI presents feed data in a compact TOON format and provides commands for reading, searching, marking, and syncing feeds, enabling autonomous agents to manage your news without a GUI. Install via Homebrew, AppImage/.deb, or direct download; the CLI can be added to an agent with the `papr-rss` skill.

An open-source system that compiles raw documents into a structured, interlinked wiki-style knowledge base using LLMs and vectorless, reasoning-based retrieval.

An AI coding agent that builds software and continuously learns a developer's specific coding style and preferences through a taste profile.

A Claude Code skill that lets an AI agent audit, fix, and continuously improve a website’s SEO, answer‑engine, generative‑engine, LLM‑knowledge, and Naver visibility. It automates the typical agency checklist, adds a content engine, and provides a measurement loop to prove results. Install via Claude’s marketplace or by cloning the repo, then run `/fire-your-seo-agency audit <site>`.

An open-source toolchain that democratizes deep learning for drug discovery, materials science, quantum chemistry, and biology.

A Python package and command-line tool for web crawling and text extraction that turns raw HTML into structured data by removing noise and boilerplate.

AssetOpsBench is an open‑source benchmark/framework for building, orchestrating, and evaluating LLM‑based AI agents that operate on industrial asset‑management data (sensors, work orders, failure modes, etc.). It provides domain‑specific MCP tool servers, several ReAct‑style agent back‑ends, 141+ realistic scenarios, and a multi‑dimensional evaluation pipeline used in KDD/AAAI/NeurIPS competitions.

SenseNova-U1 is a series of native multimodal models that unify visual understanding and generation into a single architecture, eliminating the need for separate modality adapters.

An autonomous, AI-driven penetration testing framework that orchestrates over 72 security tools to automate reconnaissance, exploitation, and reporting.

This Obsidian plugin exposes the vault via a secure, authenticated REST API and a built-in MCP server, giving scripts, browser extensions, and AI agents full CRUD, search, targeted patching, event streaming, and command execution while blocking access to the config directory.

A comprehensive framework of rules and tools for 1C:Enterprise development that enables AI agents to follow standardized coding, testing, and quality assurance workflows.

A byte-level language model designed for continuous learning that manages its own architecture and memory by paging experts between disk and VRAM, allowing it to train on modest consumer hardware.

AlphaFold 3 is an AI system for predicting the 3D structures of biomolecular interactions, helping researchers model how proteins and other molecules interact.

Gas City is a Go SDK that lets you declaratively configure and run multi‑agent coding workflows. It provides a supervisor loop, multiple runtime back‑ends (tmux, subprocess, Kubernetes, etc.), and a persistent work‑tracking store (Beads/Dolt) to orchestrate AI‑driven agents.

An AI agent framework for ESP32-series IoT devices that enables users to define device behavior through chat and execute decision-making locally on the edge.

An open-source AI gateway that provides a unified API to route requests to 1,600+ LLMs and multi-modal models, featuring automatic retries, fallbacks, and guardrails.

A knowledge management framework that enables AI agents to research, compile, and maintain structured wikis for personal projects and complex information gathering.

An interactive visualization tool that runs a live GPT-2 model in the browser to help users learn how Transformer-based models predict text.

An open-source toolkit for optimizing and deploying deep learning models across CPU, GPU, and NPU hardware to boost inference performance.

A compact, human-readable encoding of the JSON data model designed to minimize token usage in LLM prompts while maintaining high retrieval accuracy.

A collection of slides, notebooks, and code for the GPU MODE lecture series — practical, community-taught material on CUDA, PyTorch GPU optimization, and advanced GPU programming topics.