Repository index
6791 AI, ML and Web3 projects · page 63 of 68
- Self-hosted Pi Agent Runner
Run Pi coding-agent sessions in isolated remote sandboxes on your own server. Self‑hosted TypeScript service that provisions sandboxes and manages agent lifecycles via HTTP.
- Timeline Studio for MiniMax H3
ComfyUI timeline editor enabling multi-shot video creation with visual memory across scenes and angles, plus local or cloud story writing for MiniMax H3.
- Qwen-Image 2.1 Toolkit
Curated collection of Qwen-Image 2.1 checkpoints, quantized versions, prompt engines, LoRAs, and associated tooling for easy download and use.
- Search Long Documents with Jev
Python tool that pairs Jev with PageIndex to retrieve relevant passages from long PDFs and text files.
- T3-Style Herdr Layout Component
Provides a T3 Code-like layout for Herdr App with Spaces/Agents sidebar and focus mode.
- GODCODER — LOCAL AI AGENT
Open cross-platform agent for code. Your code stays on your machine, LLM keys are yours.
- Personal AI Service
blendi-remade/unreel — your personal AI‑streaming service that selects and streams videos according to your preferences in real time. The project is built on TypeScript and uses modern content generation models.
- Code Patent Analog Search
Rust tool that finds existing patents similar to your code using AI models and local search. Helps avoid duplicating inventions.
- Universal Performance Profiler
Strobe measures and explains performance of web pages, Electron apps, Bun/Node, and any process via an agent-native CLI with MCP support.
- Shizuku Apps Catalog
The project collects and showcases open-source Android apps that use the Shizuku framework to obtain system permissions without root.
- GTA IV Native on Apple Silicon
Statically recompiled GTA IV 1.0.8.0 for Mac with Metal renderer, TAA, and 100+ fixes. Build from your own game files.
- CJK-First Agent Memory Module
Python memory module optimized for Chinese, Japanese, and Korean languages, designed to integrate with mainstream AI agents.
- KVarN: KV-Cache Quantization
KVarN is a native KV-cache quantization backend for vLLM, increasing context by 3-5x, raising throughput above FP16 and maintaining FP16 accuracy without calibration, activated by one flag.
- Weight-Direction One-Step Image Synthesis
Official PyTorch implementation of the efficient CVPR 2026 one-step image synthesis method with weight direction.
- Open Super Agent
Open-source Super Agent for self-hosting — free alternative to Genspark.ai with agent orchestration, deep research, and generation of slides, tables, and images.
- Tokyo 3D Browser Model
Real-time 3D model of Tokyo built from public data, showing traffic, trains, birds, and adjustable time of day.
- Project Manager Skill
Provides ChatGPT with built‑in tools for tracking milestones, coordinating implementation, performing exact‑build QA, and storing project records.
- Java RAG Knowledge Base System
A Retrieval-Augmented Generation system for importing knowledge bases and answering questions, featuring third-party MCP integration and permission isolation.
- Model Benchmarking Platform
The project provides tools for comparing machine models as real-time controllers of engineering loops, measuring their latency, stability, and resource consumption.
- Low-VRAM Krea-2 LoRA Training
The project enables efficient Krea-2 fine-tuning via LoRA even on GPUs with limited VRAM, using mixed-precision and gradient checkpointing techniques.
- Adaptive Agent Inference Framework
EvolvingInfra dynamically optimizes inference for diverse agent workloads, adapting scheduling and resource allocation in real-time.
- AGENTSMITH: Universal AI Harness
agentsmith — lightweight core + task profiles for any AI agent, assembled with a single shell script.
- LoRA Fine-Tuning Sprint
Project accelerates LoRA fine-tuning by fixing task and hardware, and posts results to a public wall-clock leaderboard based on modified nanogpt.
- Claude Commerce Examples
Shopify implementations of Anthropic's commerce‑agents blueprint: a storefront agent and a merchant agent in Python.
- iOS App Data Manager
Manage, backup, and clean cache for iOS apps with this Objective‑C utility.
- Multiagent local engine
Tool for launching and managing multiple AI agents on one machine, allowing them to interact and complete tasks jointly.
- SMART PLAYER IN MINECRAFT
AI agent plays Minecraft by itself! Server‑side Fabric mod 1.21.3 — LLM plans actions, bot mines, builds and survives. ⭐211
- Modern Petclinic with Play React
A modern rewrite of Spring Petclinic using Play Framework and React for the vet clinic domain.
- Curated DPI Tools & Frameworks
A curated list of 200+ open-source tools, frameworks, and reference implementations for building digital public infrastructure across identity, payments, and data exchange.
- Play Framework AdminLTE Dashboard
A Scala Play framework admin dashboard built with AdminLTE, featuring modular UI components and backend integration for desktop applications.
- Windows 10 Deployment Toolkit
Automates Windows 10 installation via PowerShell scripts and creates bootable USB drives for rapid deployment.
- Play Framework 2.4 Tutorial
A structured learning project for Play Framework 2.4 featuring comprehensive wiki documentation and reference pages for JavaScript developers.
- AI Agent for Knowledge Graphs
MindAct connects Claude, Obsidian, and RAG systems to build and query community-driven knowledge graphs.
- Excel Reporting & Analysis Toolkit
Web-based Excel templates for financial reporting, budgeting, and data analysis workflows.
- Browser-Based Spreadsheet Tool
A browser-based spreadsheet application supporting tables, charts, and formulas for data analysis and management.
- Windows Product Key Extractor
WPF utility that extracts and generates Windows product keys from embedded PKEY configuration files for system recovery and deployment.
- Real-Time Operations Intelligence Dashboard
A solution accelerator providing real-time operational dashboards with anomaly detection and Microsoft Fabric data analytics for intelligent monitoring.
- Healthcare MLV Demo Template
Starter template for healthcare data agents using Fabric MLV with lake views and ontology modeling.
- DVD-R Drive Timing Calibration
Calibrates drive timing geometry for accurate DVD-R data side imaging. Python tool for optical media calibration.
- AI Learning Mini-Project
The project shows how to build an AI-driven mini-project in a week using Taro, React, and FastAPI, with question generation, instant explanations, web search, RAG base, AI illustrations, and WeChat login.
- 🤖 MOBILECLAW
Open AI agent runtime environment for Android with phone control, app automation, screen reading via VLM, and Mihomo VPN support.
- Cosmos AI Framework
This is a Python framework for training and inference of Cosmos models, providing ready-made pipelines optimized for NVIDIA GPUs and distributed computing.
- Qoder Workflow
Full end-to-end framework and high-performance OpenAI-compatible gateway for the Qoder ecosystem.
- Native macOS File Organizer
A native macOS application written in Go using the MyGo framework for managing and organizing files on shelves.
- Noumena Code TUI
Noumena Code is a terminal code editor created for the Noumena inference stack, with fast startup and deep LLM integration.
- Inertia.js Server Protocol
An ALPHA implementation of the Inertia.js server‑side protocol in Rust, providing a framework‑agnostic core and an Axum adapter for seamless integration.
- Agentic RL Tutorial & Benchmark
33-chapter Chinese tutorial from RL concepts to GRPO with TRL examples, plus reproducible discriminative-model benchmark comparing encoder, LLM-LoRA, and rule baselines.
- Uncensored Video Generator
The project provides a tool for creating high-quality video without content restrictions using AI models.
- LoRa Base Station
Open LoRa base station for Meshtastic/MeshCore on Raspberry Pi with SX1302/SX1303: passive multi-channel capture, local panel, and native forwarding.
- MEMEX: MEMORY FOR AI AGENTS
Zettelkasten memory for AI agents: sync via Markdown and Git, no vector DB. Supports Claude Code, Cursor, Copilot, etc.
- EVOQUANT: AI AGENT FOR QUANTITATIVE RESEARCH
EvoQuant — an AI agent that performs the quantitative research loop: from hypothesis generation to strategy backtesting. It operates autonomously in Python, using a multi‑agent architecture to hunt for alpha.
- MoE Inference on Consumer GPUs
Runs frontier-scale MoE models on RTX 3060 (12GB VRAM) by keeping dense layers on GPU and swapping experts from NVMe into a bounded VRAM pool.
- AI-GoldenAPP Catalog
Catalog of advanced AI applications for generating realistic images and video with minimal restrictions, high quality, and full control.
- Galahad: Blindness Research
The Galahad project studies instruction blindness in vision-language-action policies, offering diagnostics and a low-rank data correction method. It includes a paper, model, deconfounded datasets, a generator, and a measurement suite.
- SGLang Deployment Optimization
Evidence-based tool that tunes SGLang parameters in real time, reducing latency and increasing throughput.
- Motion Video Workflow Harness
Claude Code skill that sequences motion graphic and video production with step-by-step validation checks.
- Viggle-Animate H3 ComfyUI Nodes
Adds ComfyUI nodes to run the Viggle-Animate minimax H3 fine-tuned model for video generation.
- Consistent Character Node
Provides a ComfyUI custom node that loads .char files to export reference mode, LoRA adapters, and guided prompts for consistent character generation.
- Uncensored AI Generator
The project allows creating images and videos without content filters using an open diffusion model. It is available for free through a simple interface or API.
- LOOMFLOW: LIGHTWEIGHT AI WORKFLOW BUILDER
LoomFlow is a lightweight AI-native workflow builder. Describe your idea in natural language, get a visual scheme and immediately publish it as an API. Install with one click via Docker.
- Open Jev with customization
Jev implementation that can be fine-tuned on your own data. It allows quickly adapting the model to specific tasks.
- AI Short Video Pipeline
Fully automated pipeline for creating short videos with AI: FastAPI orchestration, Spring Boot gateway with fault tolerance, metrics and observability, and quality control via prompt anchoring, CLIP, and automatic AV-sync recovery.
- Goal-Flow Graph Agent
Goal-Flow is a LangGraph framework that combines workflow graphs and agent loops, translating Dify DSL into executable code and switching Dify/OpenAI protocols.
- Pure HTML Landing Kit
A collection of premium, single‑file HTML landing pages, motion components, and interactive widgets that require no frameworks or dependencies.
- System 1 Model Framework
Framework to train, test, and serve System 1 decision models on language and multimodal platforms.
- Schema-Aware SQL Access Control
Filters database schemas by user permissions before LLM queries, enforcing row-level security across Postgres, Oracle, MySQL, and SQL Server. Cuts prompt tokens by 75%.
- Tiny LLM runs inside Qwen
Karpathy's stories260K model writes stories on a vector unit compiled into Qwen3.5-4B's weights, demonstrating LLM-in-LLM NPU execution.
- Agent-Friendly Codex Images
Enables any AI agent to invoke Codex's built‑in image generation via MCP, CLI, or Skill modules. Includes ready‑made templates for Xiaohongshu posts, video covers, and Mondo posters.
- Academic PDF Figure Vectorizer
Extracts figures from academic PDFs as self-contained SVG vector graphics and editable PPTX slides, with pixel-perfect comparison and live preview.
- MCP Server for Tennis
The MCP server provides live scores, odds, and win probability to LLM agents for tennis.
- MiniMax-H3 PDD Acc in ComfyUI
Official MiniMax-H3 8-step PDD Acc LoRAs for ComfyUI support parallel head decoding, use the Euler scheduler on trained sigmas, and generate audio and video in 8 steps.
- Frontend de-identification tool
Web tool for automatically detecting and masking personal data in PDF, Word, Excel, TXT, and Markdown. Everything runs locally, data never leaves your computer.
- Crypto Accumulation Alert Bot
Scans markets for accumulation patterns and sends real-time alerts to Discord and Telegram channels.
- Canary: QA Harness for Claude
Canary is an E2E testing framework created to work with Claude Code. It records screens, console logs, network HAR files, and Playwright traces.
- Wonder-Pill: Idea Map
Claude skill that turns brainstorming into an interactive map of inverted assumptions. No answers, no ranking — just provocations.
- Brazilian Company Lookup Server
MCP server and CLI for querying Brazilian companies by CNAE and region using official Receita Federal CNPJ open data.
- NSFW Content Creation Plugin
Purge-compatible DeepSeek Harness plugin for creating adult images, voiced videos, and multi-shot films with self-hosted models, reference editing, LoRA control, and agent storyboarding.
- Windows Vector Design Alternatives Guide
Curated guide for installing Affinity Designer alternatives on Windows with step-by-step setup instructions.
- AI Image Generator
Simple Python tool for generating images and video with AI on Windows. Supports quick startup and parameter configuration through GUI.
- AI Image and Video Generator
C# library for AI image and video generation, focused on creative self-expression and content creation.
- Geometry-Based Cross-Modal Alignment
Implements shared geometry approach to align modalities without paired data, enabling unsupervised cross-modal learning for vision, language, and audio.
- MeshCore Touch Firmware
Custom MeshCore companion firmware with touch UI for multiple LoRa boards. Compatible with official MeshCore apps.
- BlackBox Mesh AI Node
Autonomous mesh node with local AI, TAK integration, and Bitcoin Cashu payments based on Meshtastic. Fully offline solution for mesh networks.
- LightVLM: VLM Optimization
Toolkit for quantization and cache optimization of vision-language models, accelerating inference and reducing memory usage.
- Imagination‑to‑Policy Data Pipeline
Converts imagined world-model rollouts into corrective datasets for training robot policies on a single GPU.
- Bilingual Urdu-English Legal RAG
RAG system for Pakistani government and legal documents with Urdu-English OCR, hybrid retrieval, and citation grounding.
- PorTAL: LoRA Adapters for LLM
PorTAL creates portable LoRA adapters, specialized for specific tasks, that are easy to transfer between different language models.
- Doom Deathmatch Claude Plugin
A Claude Code plugin that launches a Doom deathmatch session while you code, letting you play and work side‑by‑side.
- Multimodal Image Retrieval System
Practical multimodal image retrieval with intent recognition, prompt engineering, Qwen LoRA, EmbeddingGemma 2, InfoNCE contrastive learning, and RAG components.
- Autonomous AI Agent Platform
Azmx provides tools for creating, training, and deploying autonomous AI agents with full control over their behavior and flexible configuration for any task.
- ESLint Plugin for AI
The plugin adds ESLint rules that search for and warn about typical AI‑generated phrases in code and comments, helping to clean the project of artificial 'trash'.
- Lightweight Agent Personality Framework
Lightweight framework for agents with personality evolution, three-tier skill loading, and heartbeat memory system.
- Agent Coding Skills Protection
The project provides a set of quality gates that detect typical AI-generated errors in code, tests, and documentation, helping agents write reliable code.
- KALETA: CMS WITH BUILDER AND MCP
Kaleta is an open-source PHP CMS for business sites. Includes visual page builder, token design system, and Claude integration via MCP. 202 GitHub stars.
- Minimal ETL Test Data Generator
Generates minimal test datasets that provably detect broken data-mapping rules in ETL pipelines, reducing test maintenance overhead.
- ID-LoRA: Voice-Image Video Generation
Custom ComfyUI node for creating voice-and-photo videos with audiovisual identity. Supports LTX2.3 — become a director!
- LABTETHER: SMART AI AGENT FOR HOMELAB
LabTether is a Go application that combines an AI agent with a Docker environment for automated deployment and management of homelab infrastructure. It uses the MCP protocol for self‑config of services and provides CLI for controlling containers.
- On-Demand Model Alignment
Aligns LLMs instantly per request for verified users or always-on mode. Integrates with Hermes, OpenClaw, and custom harnesses.
- Data Rescue Toolkit
Rescue data from failing drives with ddrescue-style imaging, NTFS recovery including deleted files, and file carving. Runs on Windows and Linux.
- Architectural Patterns for AI Systems
Catalog of 21 architecture patterns (RAG, agents, vector DBs, inference) with a tutorial and links to open-source prototypes. In Russian and Chinese.