Trending AI projects

A WorkBuddy skill that fetches McDonald’s China menu and coupons via the MCP API, then runs a deterministic Python optimizer to find the cheapest legal combination of items, respecting budget, calorie and delivery constraints. It can be used in chat or via a command‑line demo, and includes a self‑test suite and an online interactive demo.

An AI agent framework for personal investment that provides autonomous market research, portfolio diagnosis, and cross-market strategy deduction.

A high-performance, open-source RAW photo editor and library built in Rust that features a unique command-based architecture, allowing it to be fully controlled by AI agents via MCP.

Inkloom is an early‑access service that uses custom AI models to generate fully specified, production‑ready logo designs. It breaks logo creation into explainable stages—brand analysis, typography, symbol construction, and composition—producing vector specifications rather than bitmap images. The repo contains the platform (auth, credit ledger, deployment tooling) built on Cloudflare Workers, PostgreSQL and React Router; the actual generation engine is not yet live but is slated for release.

An open-source, Rust-based reimplementation of AutoCAD that provides a professional drafting workflow and is designed to be fully controllable by AI agents via MCP.

Voicebox is an open‑source, locally‑run AI voice studio that clones voices, generates multilingual speech with eight TTS engines, provides Whisper‑based dictation, and exposes a REST/MCP API for integrating voice I/O into apps and AI agents.

A research-backed AI-writing de-flavoring skill that detects and minimally rewrites AI-style patterns using measurable indicators from controlled experiments.

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

A next-generation AI memory management system that emulates human-like cognitive processes to provide agents with dynamic, evolving long-term memory.

PentAGI is a self-hosted, multi-agent system that automates penetration testing by using AI agents to plan and execute security tools within isolated Docker sandboxes.

HarnessRouter is an open‑source Docker‑based server that unifies many AI “harnesses” (e.g., Codex, Claude Code) behind a single OpenAI‑compatible API, offering session‑persistent workspaces, streaming, file handling, and custom harness configuration.

MCP for Blender is an open‑source bridge that lets LLM assistants (Claude, Codex, Cursor, VS Code, etc.) control Blender via a socket‑based MCP server. The AI sends Python commands to create, edit, texture, and render 3‑D scenes, with optional premium AI‑generated model services. Install via a one‑line script (or Docker/pipx), enable the Blender add‑on, and start the server to work interactively from your chat client.

A powerful WebUI for few-shot voice conversion and text-to-speech that enables high-quality voice cloning and synthesis with minimal training data.

An open-source agentic development environment that provides a unified interface for multiple AI coding agents across desktop, mobile, and CLI.

Multica is a self‑hostable (or cloud) platform that lets you treat LLM‑based coding agents (Claude Code, Codex, Cursor, etc.) as teammates on a Kanban board. Agents run on a runtime you control, keep context, log every step, and hand back pull requests for human review. The stack includes a Next.js web UI, Go backend, PostgreSQL, and an agent daemon that spawns any supported CLI. It offers cost analytics, review gates, Slack/Telegram integrations, and a full CLI/API for automation. Licensed under a custom Apache‑2.0‑style license.

DeepOpen-3d-webs (Guangyu Zaojing) is a Docker-shipped, image-driven 3D scene generator: give it one reference image and a prompt, and it produces camera-trajectory scene results — color/depth animations and Gaussian point clouds — previewable in the browser. It has five tabs (Workbench, Scene Generation, Example Gallery, Task History, System Status), runs in GPU-free demo mode by default, and switches to a full LucidDreamer-based generation engine when CUDA and model weights are available. No login; anyone who can reach port 8090 can use it.

Immich is a high-performance, self-hosted photo and video management solution that brings cloud-like features—automatic backup, multi-user sharing, facial recognition, and AI-powered search—to your own hardware. It pairs a server with mobile and web apps to give you full control over your media library.

An open-source educational repository that combines a four-volume textbook series with hands-on labs, a from-scratch ML framework builder (TinyTorch), hardware kits, a simulator, and interview prep — all designed to teach the discipline of engineering reliable, scalable AI systems, not just models.

EdgeEver is an open-source, AI-native knowledge base and note-taking platform that offers a privacy-focused, self-hostable alternative to Evernote with built-in AI agent integration.

A high-performance headless browser built from scratch in Zig for AI agents and automation, offering significantly lower memory and CPU usage than Chromium.

An open-source desktop suite that installs and manages all its own local AI engines (LLMs via Ollama/LM Studio/etc., images/video via ComfyUI) behind a friendly GUI — chat, generation, coding agent, agent with tools, voice, and (in beta) computer control, all offline with an optional cloud tier.

A high-performance CLI proxy that compresses bash output for AI agents, reducing input tokens by up to 90% to lower costs and save context space.

Open WebUI is a self-hosted AI platform that provides a user-friendly interface for interacting with AI models, supporting local models via Ollama and any OpenAI-compatible API. It offers features like plugins, role-based access control, persistent memory, RAG, and more, and can run entirely offline.

A collection of AI prompts and agent skills designed to help researchers rewrite academic papers by eliminating defensive, self-doubting language and focusing on clear, persuasive arguments.

anti‑slop is a rule‑based filter for AI coding assistants that strips generic UI fluff, hype‑filled copy, and unnecessary code comments. It provides 38 mandatory rules, a delivery‑gate report, and optional skill modules (UI, copy, accessibility, etc.). Install via `npx antislop-ai` or as a plugin for agents like Claude Code, Cursor, or Codex. Use it during generation or to audit existing output, then add your own `DESIGN.md` for real design direction.

n8n is a fair-code platform for building and deploying AI agents and automated workflows, connecting over 1,500 integrations with flexible model support.

A privacy-first, open-source voice-to-text application providing local AI dictation, meeting transcription with speaker diarization, and AI-powered notes.

VisionHOPE is a visual backbone architecture that uses self-referential nested learning to dynamically evolve its memory and update rules while processing images, providing improved performance for classification and dense prediction tasks.

A collection of structured AI skills and theoretical guides that distill senior academic experience into LLM prompts to assist researchers with idea evaluation, paper writing, and peer review.

Laya is a fast, multilingual decision engine that classifies text into structured answers (choice, score, yes/no) in a single forward pass, without text generation. It uses a router to pick the best model per request and supports fine-tuning for domain-specific accuracy.

A gateway proxy that converts WorkBuddy and CodeBuddy services into standard OpenAI-compatible APIs, featuring multi-account management, automated daily tasks, and strict usage guardrails.

geo-sleuth is an agent skill that finds where a photo was taken. It turns geometry in the image—skyline, shadows, pier spacing—into search constraints against OpenStreetMap, elevation data, and satellite imagery, then returns a camera position with an error radius and an evidence image.

dsh-context is a DeepSeek Harness plugin that adds UI panels and a slash command for deep inspection of an LLM agent’s context window, token usage, cost, and file activity, with dashboards for session‑level insights, per‑turn trends, and agent‑family (fleet) monitoring.

AstrBot Launcher is a desktop application that provides a graphical interface for managing, backing up, and configuring isolated AstrBot instances and their Python environments.

An open-source image and video generation studio that lets users access 38 different AI models using their own API keys to avoid vendor lock-in.

Dify is an open‑source LLM app development platform with a visual workflow builder, RAG pipeline, autonomous agents, multi‑model support, and built‑in observability. It can be run via Docker‑Compose for self‑hosting or used as a SaaS offering.

InkOS is a Node‑based AI authoring platform for novels, short stories, scripts, interactive worlds, and translation. It provides a unified Studio web UI, terminal UI, and CLI that share a pi‑agent harness: a language model proposes actions and InkOS executes them deterministically, managing story‑wide memory, chapter snapshots, and structured state. Built‑in Skills cover writing, reviewing, cover‑image generation, and play‑world creation. It works with many LLM providers (Moonshot Kimi, Claude, Gemini, DeepSeek, etc.) via the kkaiapi aggregation layer, supports multi‑model routing, and can be called as an OpenClaw skill. Installation is a single `npm i -g @actalk/inkos` and the workflow follows create → write → review → revise → export, with robust configuration via Studio UI or environment variables.

uzu is a cross-platform inference engine for running AI models locally in your app, offering zero-latency, private, and cost-free inference with a simple API and support for multiple programming languages.

dsh‑bili‑scroll is a DSH plugin that automatically searches recent B‑site videos, reads subtitles (or locally transcribed speech), selects worthwhile items, writes a Markdown report, and can post @‑mentions for you. It runs in a self‑contained Python environment (downloaded via uv) and integrates with the bilibili‑mcp AI model for searching and subtitle processing.

Local code intelligence for developers and AI agents. Connect code, docs and infrastructure in a persistent graph, with CBM Lens, MCP tools, agent integrations and change-aware test selection.

Jasna is a free GPU-accelerated tool that detects and restores mosaic-blurred regions in videos and images. It uses a fast free model (BasicVSR++) or a slower, more detailed supporter-only model (LTX, NVIDIA only), works via GUI or command line, and requires an NVIDIA GTX 16-series+ or experimental AMD RX 7000/9000 GPU plus a disk path with only English characters.

A collection of Python scripts and design guidelines for creating publication-quality scientific figures, designed to be used both manually and as a skill for AI coding agents.

PixelRAG is a visual retrieval-augmented generation framework that renders documents as screenshots to enable AI models to search and understand visual content like tables and charts that traditional text-based RAG systems often miss.

DSHA is an Android launcher that runs the DeepSeek Harness agent framework in a full, pre-configured Ubuntu environment, enabling AI-driven automation on mobile devices without root or command-line setup.

DBX is a lightweight database management tool that supports over 100 different databases in a single 25 MB package, available as a desktop app, CLI, Docker image, and featuring a built-in AI assistant and MCP server.

Breeze TTS 2 is an open-weight, high-performance text-to-speech model that supports real-time streaming, voice cloning, and instruction-based voice design in English and Chinese.

A free, curated directory of LLM APIs that always stay free, split into first-party model providers (Aion, Cohere, Google, Mistral, Z AI) and third-party inference hosts (Cloudflare, Groq, Hugging Face, OpenRouter, etc.), each with model specs, rate limits, base URLs, and free-tier caveats like data reuse or usage caps.

An open-source LLM engineering platform for observability, prompt management, and evaluation to help teams develop and debug AI applications.