Inspired by dnhkng's RYS work, this project:
A production-ready RAG (Retrieval Augmented Generation) system for Nova Clinic that ingests medical documentation, creates vector embeddings, and answers questions using Claude AI with source citations.
Using GPT-5 (or Claude 4.5 opus, Gemini 3 pro, Llama 3.2 on Vertex AI), generate a 3D CAD model (STEP file) from a 2D CAD image.
Using GPT-5 (or Claude 4.5 opus, Gemini 3 pro, Llama 3.2 on Vertex AI), generate a 3D CAD model (STEP file) from a 2D CAD image.
为山东银座商城黄金/珠宝零售业务打造的AI Agent系统,分三期交付: 1. Phase 1: 智能营销Agent(文案/视频生成 + 网络趋势 + 线上顾客喜好) 2. Phase 2: 销售决策Agent(ERP数据接入 + 销售分析 + 顾客画像) 3. Phase 3: 产品设计Agent(结合营销+销售数据 → AI辅助产品设计)
wyh15606343555-creator__xhs-content-tool
wyh15606343555-creator__ad-spy-agent
> 基于火山引擎豆包 API 的 MCP (Model Context Protocol) 服务器,为 AI 客户端提供强大的多模态生成能力
wwchen256__chembl-query-assistant
Crypto-Geopolitical Intelligence Dashboard for the China-Japan Corridor
Context Manager is an intelligent memory management system designed to solve the context window limitations in large language model (LLM) applications. It treats context as expensive RAM and external data as pageable disk/cache, providing a "memory manager" subsystem for models.
A chat agent with tool-calling capabilities, built with FastAPI, PostgreSQL, and the OpenAI SDK.
wondersmiths__realtor-ai-agent
PDF to AI Agent Knowledge Bridge
An open-source framework for building autonomous AI agents that can execute tasks, manage resources, and interact with the real world.
wisent-ai__OpenEnv
A local voice-to-text dictation application for Windows using OpenAI Whisper. Type with your voice in any application - browsers, chat windows, code editors, and more.
> Detects, diagnoses, and fixes Oracle database problems automatically. > Drop a .md file to add new alert types — zero code changes.
whfg21746-ship-it__avto-parser
AI-assisted email coaching system for The Launchpad Incubator. Reads participant emails, drafts coaching responses in Wes's voice using GPT-4o, and sends after approval via a Streamlit dashboard.
多部门协作的 AI 舆情分析和竞品监控平台
AI人狼ゲーム — 1人のプレイヤーが8体のAIと人狼を遊ぶWebアプリケーション。
River Algorithm memory consolidation for AI applications.
Unless otherwise specified: - maximum number of layers offloaded to GPU - local models run with llama.cpp server and .gguf formats - parameter changes carried over into following tests (temperature, penalties, etc.) - * denotes non-local for comparison
A video download, transcription, translation, and encoding microservice.
LLM-powered entity extraction + hybrid RAG (vector + graph) from Griffith's Ramayana translation.
Automated Snapchat control with AI-powered actions via ADB on a real Android phone.
Voice AI productivity tool that converts speech into structured, intelligent text. Speak raw thoughts and get clean emails, PRDs, LinkedIn posts, developer comments, and more.
A working implementation of the llms.txt standard for Gies College of Business at the University of Illinois Urbana-Champaign.
Your LLM never forgets. Even in a 500-turn conversation.
A multi-agent Retrieval-Augmented Generation system built with FastAPI, Qdrant, PostgreSQL, and LLM providers (OpenAI / Anthropic / Google Gemini). Agents are orchestrated via Kahn's topological sort, enabling parallel execution where dependencies allow and can handle any type of order execution of
Encord Task Agent that uses Qwen2.5-VL-7B (via vLLM) to pre-annotate humanoid robot egocentric manipulation videos with Ego4D-style verb+noun action decomposition.
AI Emergency Response System — Automated ePCR for Paramedics.
Open-source real-time voice-to-text desktop app for Windows and macOS. Dictate into any application with a global hotkey. Supports local and cloud STT engines, offline translation, and AI-powered text polishing.
A lightweight OpenClaw alternative. One base class, 24 built-in tools + MCP extensibility, extend in 15 lines.
Vijil Dome is a fast, lightweight, and highly configurable library for adding runtime guardrails to your AI agents. It combines top open-source LLM safety tools with proprietary Vijil models to detect and respond to unsafe content — with built-in support for observability, tracing, and popular a
> New version available: Nous Desktop — native macOS app with web UI and local model support. This repo is the original Colab/Telegram version.
Extract transcripts from video URLs across multiple platforms using Whisper AI.
자율주행 주행 데이터를 한국 도로교통법(도로교통법)에 따라 자동 평가하는 법규 준수 엔진.
Version: 2.0.0-alpha Release Status: In Development Author: Takeshi Fujishita
veraymw__quinn-learning-service
llm-box brings the power of large language models to your command line. Instead of just listing files, it describes what they do. Instead of dumping file contents, it explains them. Instead of searching by filename, it understands what you're looking for.
Shared LLM abstraction layer for chess coaching applications.
1. Download from python.org → Python 3.12.7: https://www.python.org/downloads/release/python-3127/
Windows-приложение для голосового ввода текста. Зажмите горячую клавишу — говорите — текст появляется в любом приложении в реальном времени.
A high-performance FastAPI service for searching and analyzing Board of Veterans' Appeals (BVA) decisions. This API provides real-time access to BVA decisions without requiring a database, utilizing the USA.gov search API and web scraping.
Automated contamination detection for SWE-bench benchmarks. Identifies tasks where gold patches or fail-to-pass (F2P) tests exceed the problem description, producing unfair evaluation criteria for software engineering agents. Features a dual taxonomy classification system (v3), LLM-primary traje
Telegram kanallar uchun avtomatik post yaratish va joylashtirish boti.
Локальный веб-инструмент для анализа Instagram-блогеров по брифу.
> Hi everyone! The original version of PaperBanana is already open-sourced under Google-Research as PaperVizAgent. This repository forked the content of that repo and aims to keep evolving toward better support for academic paper illustration—thoug