Projects approved for live demonstration • 10 projects
Codechimp is the first AI-native project management platform built for vibe coding, seamlessly integrating humans and AI agents. It enables scalable vibe coding across teams, organizations, and the entire development lifecycle.
SmartLegalPredictor is an AI-powered system that automates the extraction, analysis, and management of legal case information through intelligent conversation and document parsing. It leverages NLWeb for NLP and web-based legal research, and Weaviate as a vector database for semantic search and comprehensive legal knowledge retrieval, enhancing legal professionals' efficiency.
This AI dream interpreter uses machine learning to expose the subconscious meanings behind your dreams, explaining why seemingly embarrassing dream scenarios might carry deeper messages. It aims to provide a modern alternative to traditional dream interpretation books.
React, Next.JS, OpenAI API, Tailwind, Shadcn, Lucide
TimberGem is an intelligent copilot for the construction industry that transforms static building documents into a dynamic, queryable digital twin. It enhances efficiency for general contractors and estimators by integrating a manual-first knowledge graph interface with AI automation, thereby improving accuracy, reducing time spent on takeoffs, and ensuring traceability.
This AI-native project uses Weaviate, GraphRAG, and NLWeb to make global treaties and policy clauses searchable, comparable, and understandable. It transforms dense legal documents into actionable insights through semantic search, graph-based reasoning, and natural language interaction, building a dynamic knowledge graph for comparative analysis.
This MCP Server enhances predictability and repeatability by saving reasoning to logic. A Loom demo is available for further details.
CodeIntegrity.ai builds guardrails for agentic AI systems, focusing on making multi-tool AI safer and more reliable. They've introduced a new feature to tackle the "lethal trifecta" problem in MCPs, using security taint flow analysis to track data movement and surface risks.
MCP, React
Mycelian Memory is an open-source framework that provides simple, reliable, and cost-effective long-term memory and context to AI Agents, organizing information in immutable timelines for high-precision recall. This framework helps AI agents persist their working context and memories, overcoming the limitation of forgetting information when a session ends.
PostgreSql, OpenRouter, used Claude Code, Codex and Cusor to build the project before today, GraphRAG [will experiment]
GraphRAG is an intelligent assistant that unlocks insights from data, with potential applications like powering a SynthBio Research Assistant for drug development. It enhances ride-sharing apps by providing explainable, context-rich answers for customer analytics, pricing intelligence, and route optimization.
This project automates the essential "Keep-the-lights-on" exercise for teams with on-call rotations, saving engineering hours by reducing manual touch points. It aims to streamline periodic production issue triage and action item management for always-on services.
Community submissions • 46 projects
This project involves building an AI chatbot that recommends courses from a database of over 5000 UW courses. The courses were ingested into a Weaviate collection, and a script using Weaviate SDK's RAG feature generates course recommendations based on user queries.
Convex
This project helps users develop complete stories from a single idea using AI agents that guide them through character building, screenplay writing, and plot design. It includes features like Plot Diagrams and an AI Bounce Board for idea validation, making story creation accessible without prior experience.
The "Digital Consciousness" agent is an AI service that unifies an individual's fragmented digital identity into a single, intelligent knowledge base. It creates a persistent, self-evolving partner that understands a user's world and provides proactive assistance based on their unique life context, transforming human-computer interaction.
LLM, GitHub Codespace,
Best Ride is an app that compares ride services from various providers like Uber, Lyft, Hertz, and Enterprise to find the best ride. It uses GraphRAG for advanced features like customer data insights, dynamic pricing analysis, and driver route optimization to provide explainable and context-rich answers.
Card Witch is a secure, subscription-based financial companion that helps users manage all their payment methods in one platform. It recommends the best payment method for purchases and optimizes financial products and spending habits, while maintaining privacy and transparency.
This is a voice-activated AI medical triage system that analyzes symptoms to detect medical emergencies and provide health guidance. It currently answers basic information for MVP and is planned to use GraphRAG in the long run.
PainPointPal is an AI-powered agent that automates user research using GraphRAG and NLWeb. It collects, analyzes, and synthesizes user feedback from multiple sources to identify pain points and provide actionable insights for companies.
Samvidhaan.ai provides solo practitioners in Indian legal practice with AI-powered document intelligence, democratizing technology previously exclusive to larger firms. It aims to empower individual lawyers with advanced tools for their practice.
ChatGPT (for ideation), Claude (for making sure ChatGPT didn't hallucinate)
This GraphRAG-powered intelligence engine transforms SEC filings by automatically extracting, connecting, and distilling insights relevant to a user's portfolio. It helps users spot risks, uncover opportunities, and make faster decisions by identifying relationships across companies, sectors, and risk factors.
SAUL is a GraphRAG-powered platform for multinational corporate counsel/risk officers to run dynamic, traceable compliance audits and risk scenarios. Its goal is to reduce outside counsel spend and compress scenario-planning cycles, targeting a legal tech market forecast to reach ~$64 billion by 2032.
Link Reel is a web application that generates animated branded post-roll videos based on user identity and promotional links. These videos are exclusively used within the Bumpr iOS app for paid subscribers, offering a demo-first, login-gated experience for configuring and previewing video styles.
This description helps college students and high-schoolers make informed career decisions.
memeED is a tool that generates kid-safe hooks and mini-quizzes for any topic, reducing lesson prep time significantly. It uses a curriculum graph and vector library to ensure on-topic content and reusability, leading to engaged classes.
React, OpenAI, Express
Ashley is an AI-powered calendar assistant that actively manages your time by prioritizing, rescheduling, and negotiating events. Built with GitHub Copilot, Ashley continuously evolves with new features, effectively turning your calendar into a Chief of Staff.
This service uses AI agents to provide real-time alerts and suggest mitigations for business disruptions like strikes or factory fires. It aims to notify users of events, show their impact, and allow for risk mitigation with a single click.
This AI-native platform helps independent developers and small studios create clear proposals and accurate cost estimates for AI projects. It addresses the chaos in AI project scoping by providing predictable pricing, timelines, and transparent cost breakdowns for both developers and business buyers.
This project modernizes a 4000-year-old Chinese divination system using technology and Graph RAG to make its philosophy understandable to a Western audience. By communicating in culturally relevant ways, the goal is to reduce misunderstandings and improve the communication of complex philosophical ideas.
StudySpark helps students find explanations at the right level of detail by matching questions to semantically similar solved examples and generating graded hints. It prioritizes privacy by avoiding PII storage and sanitizing inputs before processing.
ResumeRAG solves inefficient resume filtering by using a graph-based approach with GraphRAG to understand relationships between skills and experiences, allowing natural language queries for accurate talent discovery. It was implemented with Python, GraphRAG, FastAPI, and React, and future plans include enabling users to upload their own resume databases.
This tool helps building construction engineers by using natural language search to quickly find information in codebooks. It presents the information in an easily understandable way, speeding up their work cycle.
Mytreya is an AI spiritual guide leveraging GraphRAG to synthesize esoteric knowledge from various cultures, offering personalized and verifiable spiritual guidance. This AI agent addresses the growing demand for personalized spiritual guidance, AI-powered wellness solutions, and mental health support within a large market opportunity.
MyChart offers financial advisors and their clients features to upload files. These files are parsed by GraphRAG to provide insights and portfolio details, which can then be addressed via a Natural Language Web chat.
Open AI
Utilize a Large Language Model (LLM) to identify the most sustainable toy options available on Amazon.
This project is a wrapper around LLM Studio to create bedtime stories for kids. There are future plans to add image generation to the project.
Constellate is a project designed as a research aide, helping users connect information to make informed decisions. Currently, its scope is limited to assisting users in finding houses in the US, with plans for future expansion.
Oikyo is developing a user-friendly, cross-platform application to simplify the fine-tuning of LLM models, making it accessible even without deep technical expertise. This privacy-first tool allows for local, secure, and cost-effective model training, offering comprehensive features from data preparation to deployment, akin to a "Postman for fine-tuning LLM models."
This platform offers lifelong personal contacts management and relationship tracking, consolidating all interactions from various platforms into one secure source. It is a subscription-based service with cloud storage and advanced functionality upsells, including a mobile version with offline capabilities for daily tasks.
The 15-year Macallan
Be A Good Friend is an AI agent that helps users re-connect with contacts they haven't spoken to in a while. The agent identifies and helps users re-engage with these contacts.
ChatPMS is a revolutionary property management system that uses natural language commands to automate hotel operations, eliminating complex UI navigation and reducing time-consuming manual tasks. It streamlines various processes including reservations, check-ins, revenue management, and housekeeping, ultimately enhancing efficiency and guest experience.
The GraphRAG Knowledge Assistant is an AI-powered tool combining GraphRAG with LLM technology for intelligent document analysis and multi-hop reasoning. It features fast LLM inference via Groq API, interactive knowledge graphs, and a professional web interface.
Test Project2
Recruiters often waste time on irrelevant matches or miss strong candidates entirely, resulting in slower hiring and overlooked talent. ResumeRAG tackles the broken resume filtering process by moving beyond traditional keyword search to a graph-based approach using GraphRAG that actually understands relationships between skills, roles, and experiences. Our solution builds a knowledge graph of resumes using GraphRAG, then lets users query it in natural language to discover and visualize relevant information. We implemented this with Python, GraphRAG, FastAPI, and React, layering LLM-driven context and visualization on top to make resume search more accurate, transparent, and useful. What’s next for ResumeRAG would be allowing users to upload their own resume database, allowing them to clearly and efficiently identify talent from within their own hiring processes. Demo: https://www.loom.com/share/bc805acf6ecd405c9f045a7fcc7d8749?sid=bc2f7d1a-dcab-4018-a4f2-4b9519726ed3
This system processes video and audio to generate follow-up content like posts, articles, summaries, and comprehension tests. It is ideal for university lectures and business meetings.
This platform teaches kids to trade in the stock market using virtual money and real stock data. It features an intuitive, chat-based interface that allows kids to buy, sell, and research stocks.
This conversational AI tool provides personalized academic guidance to students, helping them navigate complex university catalogs by answering natural language questions. It serves as a powerful proof of concept for a tool that can save students hours of research, applying the latest AI and vector search technologies.
Flixly is a full-stack web application that allows users to search for movies and TV series and compare their availability across multiple streaming platforms. It provides details on availability type, pricing, video quality, region, and direct links to watch.
TMDB, UTelly
This description focuses on the tools used to correlate code and documentation.
Expert network is a system that allows users to connect with experts in various fields. It facilitates knowledge sharing and collaboration.
This prototype is a FastAPI-based demo for small organizations to discover relevant grants/RFPs and produce a short checklist and a draft cover letter. It combines web signals and document-backed evidence to surface funding opportunities and quickly generate next-step checklists and draft cover letters.
The VeneoHost Agent is an AI-native representative for a single venue with rentable meeting spaces, designed to assist prospective renters and streamline operations for short-term rental property owners. It provides authoritative answers by combining data from Veneo's database, vector store, and the venue's public website, with a modular and integratable design.
planning to use Weaviate :)
Recruiters often waste time on irrelevant matches or miss strong candidates entirely, resulting in slower hiring and overlooked talent. ResumeRAG tackles the broken resume filtering process by moving beyond traditional keyword search to a graph-based approach using graphRAG that actually understands relationships between skills, roles, and experiences. Our solution builds a knowledge graph of resumes using GraphRAG, then lets users query it in natural language to discover and visualize relevant information. We implemented this with Python, GraphRAG, FastAPI, and React, layering LLM-driven context and visualization on top to make resume search more accurate, transparent, and useful. What’s next for ResumeRAG would be allowing users to upload their own resume database, allowing them to clearly and efficiently identify talent from within their own hiring processes. Demo: https://www.loom.com/share/bc805acf6ecd405c9f045a7fcc7d8749?sid=bc2f7d1a-dcab-4018-a4f2-4b9519726ed3
This tracks the time spent in money.
IT uses Weaviate for vector Search. Loom Link
The user is attempting to complete a demo using Docker from a third-party WebXR React tutorial found on GitHub. They note that the repository is not their own.
Recruiters often waste time on irrelevant matches or miss strong candidates entirely, resulting in slower hiring and overlooked talent. ResumeRAG tackles the broken resume filtering process by moving beyond traditional keyword search to a graph-based approach using GraphRAG that actually understands relationships between skills, roles, and experiences. Our solution builds a knowledge graph of resumes using GraphRAG, then lets users query it in natural language to discover and visualize relevant information. We implemented this with Python, GraphRAG, FastAPI, and React, layering LLM-driven context and visualization on top to make resume search more accurate, transparent, and useful. What’s next for ResumeRAG would be allowing users to upload their own resume database, allowing them to clearly and efficiently identify talent from within their own hiring processes. Demo: https://www.loom.com/share/bc805acf6ecd405c9f045a7fcc7d8749?sid=bc2f7d1a-dcab-4018-a4f2-4b9519726ed3
Recruiters often waste time on irrelevant matches or miss strong candidates entirely, resulting in slower hiring and overlooked talent. ResumeRAG tackles the broken resume filtering process by moving beyond traditional keyword search to a graph-based approach using GraphRAG that actually understands relationships between skills, roles, and experiences. Our solution builds a knowledge graph of resumes using GraphRAG, then lets users query it in natural language to discover and visualize relevant information. We implemented this with Python, GraphRAG, FastAPI, and React, layering LLM-driven context and visualization on top to make resume search more accurate, transparent, and useful. What’s next for ResumeRAG would be allowing users to upload their own resume database, allowing them to clearly and efficiently identify talent from within their own hiring processes. Demo: https://www.loom.com/share/bc805acf6ecd405c9f045a7fcc7d8749?sid=bc2f7d1a-dcab-4018-a4f2-4b9519726ed3