Anthropic AI Hackathon Summary

Hey there! We are the Digital Support Gurus, a team of tech enthusiasts participating in the thrilling LabLab's Anthropic AI Hackathon. As the Team Lead, I am incredibly proud to introduce our groundbreaking creation—a powerful Ecommerce Customer Support Chatbot that will revolutionize the way businesses engage with their customers. Our journey began with a deep dive into the world of language models, where we discovered the incredible potential of LangChain, a Python framework tailored for developing applications powered by language models. Harnessing the capabilities of two exceptional models, we embarked on a mission to create an unparalleled customer support experience. At the heart of our chatbot lies the OpenAI model, a dynamic entity that diligently tracks chat history and skillfully poses follow-up questions. This ensures that every conversation is seamless and contextually relevant, building a strong foundation for an exceptional customer experience. But our true hero is Anthropic's Claude Model, armed with an impressive 100k context window. It serves as the ultimate shopping assistant, providing unrivaled accuracy and delivering the best possible results. We, the Digital Support Gurus, are here to revolutionize the way you engage with Ecommerce—one conversation at a time. Prepare to be amazed as we redefine the very essence of customer support. Get ready for an extraordinary journey where technology meets innovation and customer satisfaction reaches new heights. Welcome to the future of Ecommerce support!

Digital Support Gurus

In today's fast-paced product development landscape, product leaders play a critical role in ensuring efficient collaboration between designers and engineers. However, creating Jira tickets can be a time-consuming task that demands a deep understanding of various domains, from business and engineering to design and external resources. That's where CreateTickets comes in. Our platform is designed to empower product leaders by streamlining the ticket creation process and providing them with structured, easy-to-read tickets that encompass all the necessary context. With CreateTickets, product leaders can now create comprehensive Jira tickets in less than 2 minutes per ticket, saving valuable time and allowing them to focus on higher-level strategic initiatives. CreateTickets goes beyond simply creating tickets; it acts as a valuable assistant, ensuring that product leaders have the peace of mind they need to drive successful product development cycles. By automating the collection and organization of vital context information, our platform eliminates the need for manual data gathering, reducing the risk of errors and miscommunication. With CreateTickets, product leaders can easily communicate their vision and requirements to designers and engineers, providing them with a clear goal to work towards. By facilitating effective collaboration and minimizing ambiguity, CreateTickets enhances productivity and enables teams to deliver high-quality products more efficiently. Join the growing community of product leaders who have embraced CreateTickets and experience the power of streamlined ticket creation. Focus on driving innovation, while we take care of the administrative burdens. Try CreateTickets today and unlock your true potential as a product leader.

Unity Team

Philosophy is a notoriously difficult genre for casual readers to breach. It is a body of knowledge spanning over centuries, which unfortunately brings with it all the archaic, outdated language of yore. Yet, contained within these dusty old tombs are precious concepts and theories that have shaped the underlying fibres of our culture. My question was thus: How can we bring these ideas into the modern age, while still preserving the author’s original words? Philaude (short for “philosophiclaude”) is an AI-powered reading assistant that aims to lower these barriers of entry. This project uses Claude AI’s language processing and large context capacity to make philosophy classics accessible for modern readers. The web app includes an in-page EPUB reader alongside a chat window, where users can supplement their reading page-by-page through questions and helpful discussions with the AI. With a digital bookshelf full of public domain works, users can save their progress in the database and continue from where they left off. Targeting laypeople and casual genre enthusiasts, this approach offers real-time support for understanding complex concepts and historical references without losing the original author's voice. Future features envision expanded reader functionality, cross-book concept connections, and AI customizability to suit diverse user experience levels.

meechmeal

Py Interpreter is a game-changing, AI-driven data analytics platform that is designed to transform the way businesses comprehend and utilize their data. Powered by Claude, an intelligent AI assistant, Py Interpreter accepts file uploads, conducts sophisticated data analysis in real time, and produces actionable insights through dynamic graphs and charts based on your business queries. The platform significantly simplifies data analysis, empowering all team members to make data-driven decisions, irrespective of their technical prowess. By eliminating the need for extensive coding skills, Py Interpreter democratizes the realm of data science, making it an integral tool for any business looking to thrive in today's data-driven world. With its user-friendly interface and capability to manage intricate analytical tasks, Py Interpreter is not merely a software solution, but a valuable partner in your business intelligence strategy. Privacy is also a paramount concern, as all data processing is executed within isolated, secure containers, leaving no remnants post-session. Leverage the power of AI with Py Interpreter and unlock the full potential of your business data. Whether it's predictive analytics, trend identification, or simply gaining a clearer understanding of your operations, Py Interpreter is your key to a smarter, more efficient, data-powered future.

Entropic Alignment

Claude Data Protection Assistant is service for Data Protection Officers and Data Protection Leads to query the UK GDPR (UK Data Protection Act 2018) with a chatbot and have it provide advice on specific use cases. Claude DPA makes use of Claude's 100k context to read the contents of the Data Protection Act, and respond to any queries that the user has while being able to reference the specific articles where its information is sourced. With the added ability to reference existing text provided via prompts and user documents, Claude can better accurately help with solutions. With an OCR Plugin included, you can upload documents for text extraction and plug it right into Claude. Additional features be the ability to review redacted documentation and identify any information that has not been redacted. The scope is currently limited to the UK GDPR, but we plan to expand it to support the European GDPR and other similar legislation.

ChatDPO

Introducing "Cruft Collector," a revolutionary browser extension designed to streamline your browsing experience on HackerNews. Tired of sifting through endless threads and irrelevant posts? With Cruft Collector, you can bid farewell to the noise and focus on what truly matters. This powerful tool harnesses advanced algorithms to intelligently filter out low-quality or repetitive content, leaving you with a refined selection of insightful discussions, thought-provoking articles, and meaningful interactions. Save precious time and energy as Cruft Collector seamlessly sorts through the clutter, ensuring that your HackerNews feed is tailored to your interests and preferences. Upgrade your browsing experience with Cruft Collector and reclaim control over your online journey.

Acadia

DocuGPT is an AI-powered conversational interface that can help users with their queries and provide them with relevant information. Essentially is is a chatbot for documentation, that can help users understand the features and functionalities of a product or service. It can provide users with step-by-step instructions on how to use the product or service, answer their questions, and clear any doubts they may have. It can be especially useful for beginners who may not be familiar with the product or service and need more guidance. Key Features: 1- Interactive Chat Interface: DocuGPT welcomes you with a friendly chat interface that invites you to explore your tool's documentation effortlessly. No more struggling with complex search queries or sifting through lengthy documents. Simply type in your questions or concerns, and DocuGPT will provide accurate and relevant responses in real-time. 2- Beginner-Friendly Guidance: As a beginner, learning a new tool can be daunting. DocuGPT understands this and offers tailored guidance to help you get started. From explaining basic concepts to providing step-by-step tutorials and examples, DocuGPT ensures that you have a strong foundation, empowering you to make progress with confidence. DocuGPT revolutionizes the way users interact with documentation. It transforms the often-overwhelming experience into an engaging and interactive conversation, making knowledge accessible to everyone. With DocuGPT by your side, you'll discover newfound confidence in your abilities, accelerate your learning process, and unlock the full potential of your chosen tool. Experience the power of DocuGPT today and embark on a documentation journey like never before!

Antematter Labs

Effectively, the integration of AI systems into a todo-list and task manager, to allow human-style communication and encouragement when completing tasks. Focused on the social aspect of learning and behavior management, by incorporating human-style conversations, notifications and pings as encouragement, it aims to turn todo list style apps, which for some people, specifically and especially those with neurodivergent traits, may be difficult to get started or stay consistent with. In the same vein as going to the library to study, either with friends or just to be in a working environment, having a human style interaction to help encourage to get started, evaluate why tasks aren't getting done, and to recommend potential improvements to tasks to help you work better, all while logging longterm information in conjunction with ChromaDB, would allow for truly personal conversations and notes for improvement, helping users to keep up with their tasks both short-term and long-term.

newbattle

Our aim is to assist the new generation of Indians in selecting the most suitable credit cards and accessing 24/7 customer service for their financial products. We strive to simplify the credit card application process, provide comprehensive research capabilities, and offer reliable customer support. To achieve this, we will build a robust system using advanced AI and foundational models like Claude. This technology will continuously update and analyze the ever-changing financial landscape, allowing us to provide up-to-date information and personalized recommendations to our customers. Our ultimate goal is to make credit card selection effortless and ensure that individuals have the necessary support throughout their financial journey. By leveraging cutting-edge AI and a user-centric approach, we aim to empower people to make informed decisions and navigate the complexities of credit cards with ease. In addition to our current focus on credit cards, we have a vision to expand our services to encompass other complex financial instruments in the future. We aspire to provide real-time support and assistance for these instruments as well. By leveraging our advanced technology and expertise, we aim to offer a comprehensive range of financial products and services. Our goal is to become a one-stop platform where individuals can confidently explore and understand various financial instruments, receive personalized recommendations, and access real-time support throughout their financial journey.By expanding our offerings beyond credit cards, we aim to empower individuals to navigate a wide range of financial options with ease and confidence.

Credit Card Banking Help Bot

The purpose of this AI chatbot is to address many of the challenges companies face in providing fast, effective customer support. With call volumes increasing and customers expecting quicker responses and more self-service options, an AI chatbot is a very efficient solution that will see increased usage as time goes on. We developed the chatbot using Anthropic's Claude to provide instant responses to common customer queries 24 hours a day via text. For more complex interactions that require human judgment or access to account information, Claude can provide the customer with appropiate contact information so they can interact with a human agent. Our goal is to decrease wait times, reduce call volumes, and increase satisfaction rates for customer service organizations. By implementing Claude, companies can optimize resources while offering convenience and quick resolution of basic issues.

The Goons

AI-powered Personal Tutor: Vocava leverages state-of-the-art Large Language Models to act as a personalized language tutor. This AI can adjust its teaching strategy according to user's fluency and interests, making each learning session tailored and efficient. - Immersive Learning: Unlike traditional language apps that focus on vocabulary and grammar, Vocava focuses on creating immersive, context-based learning experiences. This mimics how people naturally acquire languages, making learning more intuitive and enjoyable. - Language Translation and Conversation Practice: Vocava offers a translation module with added features like part-of-speech tagging and explanations. Moreover, users can engage in conversation with the AI tutor in the Chatterbox module, practicing their speaking and listening skills. - Storytelling and Reading Comprehension: The Storytime module presents learners with stories in their target language and offers comprehension questions, reinforcing understanding in an entertaining way. - Culture Corner: Vocava goes beyond language learning, offering insights into the culture and traditions of different regions. This helps users understand the context of the language and adds richness to the learning experience. - Learning Through Games: Vocava's Arcade module presents a series of games that teach language in a fun and engaging manner. From Pictionary and MadLibs to Jeopardy, learning becomes a delightful activity rather than a tedious chore. - Dynamic Vocabulary Learning: The Playground module allows learners to generate new vocabulary and phrases, save known phrases, and review them. All these phrases are embedded in a vector database for future reference. - Analytics Dashboard: Vocava offers a comprehensive dashboard to track learner's progress over time, making it easy to see improvements and identify areas for focus. - Newsfeed: Users can access real-time content in their target language, practicing their skills with actual, relevant information.

The Irrelevant Elephant

An AI-powered story generator that combines biodiversity data and animal behavior science research to create engaging narratives about wildlife. The goal is to produce engaging and educational narratives that can raise awareness of biodiversity issues and inspire conservation efforts. The system uses LLM to generate stories from supplied data sources. It integrates data from biodiversity monitoring of animals in Upemba National Park, scientific research on animal behavior, and examples of wildlife stories for different audience levels. After initial training, the AI agent can generate customized narratives for specific purposes. There are three main settings: Kids Tale: The AI tells simple stories suitable for children to teach basic biodiversity concepts in an entertaining way. (Outside scope of MVP) Main Tale: It's aimed at the general public and focused on impact, highlighting the importance of wildlife, our dependence and the human impact on it. The stories have an educational element while remaining engaging and are oriented towards a CTA. Research Tale: It’s aimed at a technical public, the AI analyzes the combined data to produce insights that can assist park managers, biologists and researchers with biomonitoring efforts and conservation strategies. (Outside scope of MVP) The narratives generated can be the perfect substrate to develop other projects, some examples: Communication Campaigns for Social Media - Main tale Gaming Assets - Main tale. Ex 1. Kids books/games - Kids tale Biomonitoring support - Research Rale

Tale Protocol

It supports using plugins to call external resources (such as internet search and knowledge base retrieval) and modifying the results returned (generating images and code, etc.) The powerful AI capabilities are used as the core CPU of the system for data analysis and task processing. Before sending messages to the AI, interception can be done to add additional information such as relevant knowledge base information (such as hwchase17/langchain), network search information, and even adjusting the request data to support multi-modal recognition. For example, image information can be parsed by other tools and analyzed by Claude after obtaining the corresponding information. After the AI responds to the question, the response result can also be processed again, such as using Stable Diffusion for image drawing or using Gmail's API for email sending. To enhance the flexibility of the implementation, Hooks are used to achieve corresponding capabilities. Before and after the user's information is handed over to Claude for processing, chat information is passed to plugins for processing. The pre-processor can modify the user's question to add context information or even directly modify the question. After that, the pre-processor submits the processed information to Claude in a specific format. Once Claude returns the information, the post-processor can analyze and process the response or call APIs from other platforms to process tasks. In this case, we need to adjust the Prompt in the pre-processor so that Claude returns Action-formatted data. { "type": "createMail", "payload": { "content": "xxx", "target": "[email protected]", "platform": "google" } }

Ai Coder

Thirdshift: Turn your cold leads hot overnight with a CRM AI integration. Walk in and hit the ground selling with realtime, relevant customer insights from stored customer data and analysis. Match Leads to Reps, Get Insights, and schedule calls, all from one platform. An AI-powered framework that integrates with your CRM platform, curating and enriching your data while you sleep. Using our custom built prompts and Claude's 100k token limit, ThirdShift can help you analyze large segments of data to derive unique insights. Trends, analysis, and advice in a a holistic, plain English data view.

AgentForge

An AI coding plugin for the Jetbrains suite of IDE's that uses Claude for it's backend. Generate intelligent code on the fly, capable of using all the files in your project as context due to large extreme context windows.

lowtierdevteam

This is a life simulation game that simulates the most important events that occur each year in a player's life. Players can customize their character's attributes and talents. CLAUDE AI, considering the style of the game, predicts the character's life. The game style can be close to reality or depict a fantastical future or past. The motivation behind creating this game is that people's attention is often focused on the near future, and humans are inherently not very adept at long-term planning. This game extends the perspective of viewing things to the span of an entire life. The backend of this project is Flask and the frontend is Vue, with most of the development carried out by me.

life unlimited

Swift Search is a YouTube video summarizer that revolutionizes the way we consume online content. This tool harnesses the power of advanced machine learning algorithms to analyze and condense lengthy YouTube videos into concise summaries, saving users valuable time and effort. Swift Search intelligently identifies key moments, highlights, and key points in the video, providing users with a comprehensive overview of the content within seconds. With its intuitive interface and seamless integration, Swift Search empowers users to stay informed, discover relevant information, and make the most out of their video-watching experience. Say goodbye to lengthy videos and hello to Swift Search's efficient and streamlined video summarization capabilities.

Codexai

This application utilizes Claude's enhanced 100k token capabilities to efficiently and accurately analyze vast files of proposed legislative bills. Its primary focus is on the discovery and elucidation of anti-Trans policies buried within complex legal documents. The app enables users to effortlessly dissect 100-page bills, providing concise summaries and critical analyses. The main objective is to enhance activism’s accessibility. While many individuals care deeply about combating anti-Trans policies, they often lack the time and resources to navigate through numerous proposed legislations. This tool empowers individuals and activist groups by streamlining the process, freeing them to concentrate their efforts on driving tangible change. Immense effort has been dedicated to engineering ClaudeJD's prompts and designing its user interface. By employing well-crafted prompts, the application delivers reliable, accurate, and concise responses, minimizing any potential inaccuracies (hallucinations). The user-friendly interface ensures an intuitive, repeatable, and enjoyable experience. Future plans include implementing features such as memory retention, sorting capabilities, and batch runs, among others. This project holds tremendous potential for expansion into new and promising domains.

ClaudeJD

SmartSnap is our mobile application (already publicly available) that uses AI technologies to help students with their school homework. The user takes picture or types in a text of the problem, our application analyses it and provides step-by-step solutions using generative AI technologies. We are building separate AI models for different subjects and plan to provide different learning experiences for each. Currently, we have a specialised model for maths and a more general one for other subjects. For the Anthropic AI hackathon, we decided to use Claude API for a new use case - helping students with text comprehension tasks. We created two endpoints: one for a summarisation task, the other for Q&A in the given text context. We integrated these endpoints into our application with a new interaction screen. Thanks to Claude AI, the user is able to scan text from a reading comprehension test, immediately receive a summary, and then "chat with text". Anthropic AI's Claude AI is an ideal solution for text comprehension tasks, primarily due to its ability to process large contexts, effectively handling up to ~70k words. In addition, the AI is designed to be helpful, honest and harmless. Maintaining the honesty of information and causing no harm is well suited to educational applications where accurate and safe answers are crucial.

Vantage Wave

Deep Dive is an innovative academic exploration tool, designed to bridge the gap between disciplines. Created for a hackathon, this web app integrates the Claude API—an advanced Large Language Model, the Mathpix API for math interpretation, and the intuitive Streamlit platform. Deep Dive's unique feature is its ability to convert PDFs into LaTeX format using Mathpix API. This function enables the understanding of mathematical expressions, tables, and formulas, creating a game-changing experience in academic paper navigation. The heart of Deep Dive is its customizable depth of understanding slider. Users from all backgrounds can select a comprehension level ranging from a basic understanding (level 1) to a detailed, technical understanding (level 3). Consequently, a computer science major can grasp a biology paper, and vice versa. Once users input their paper and select the comprehension level, Deep Dive employs Mathpix and Claude APIs to generate a detailed summary tailored to the chosen depth of understanding. Users can ask questions about the paper, with the app producing answers matching the comprehension level. For brevity, responses can be condensed via a checkbox. Deep Dive also includes an interactive multiple-choice quiz designed using the Claude API. These quizzes adapt to the chosen depth level, providing instant feedback and explanations for correct answers. In essence, Deep Dive democratizes academic content, making interdisciplinary learning more accessible and engaging through its mathematical comprehension, customizable summaries, Q&A, and quizzes. It's a milestone in academic research tools, transforming learners into active participants in the global knowledge community.

AI

The proposed tool aims to assist students in learning English by simplifying the process of checking word definitions within the context of sentences. Students often find it time-consuming to look through various definitions and discern which applies to the sentence they are analyzing. Our tool intends to solve this issue by providing an easy-to-use platform for students to input their text and receive accurate definitions of the words. By streamlining the learning process, students have more time to focus on the actual comprehension of the sentence and improve their vocabulary. Our tool will be a valuable resource for both language learners and educators looking to facilitate English language acquisition.

ChaCha

In our rapidly evolving society, each commercial building's financial life cycle undergoes rigorous management, auditing, and examination by a complex network of professionals. This chain includes finance managers at the building, owner, joint venture, lender, and prospective buyer levels, among others. A conservative estimate suggests that, in the United States alone, approximately $27 million is expended monthly on labor hours solely for the interpretation of financial statements. AssetGPT is poised to address and resolve this issue. AssetGPT, an AI-driven tool, is primed to revolutionize real estate asset management. It streamlines the intricate tasks associated with managing financial statements, providing much-needed solutions to the laborious and daunting processes that financial managers encounter daily. Empowered by advanced AI capabilities, AssetGPT offers lucid summaries of financial reports, conducts comprehensive audits, and provides illuminating insights into the embedded data. Leveraging our pioneering technology, challenging financial procedures are transformed into seamless, automated workflows, thereby enhancing both efficiency and accuracy. In the future, AssetGPT is set to expand to accommodate additional document types, including Rent Rolls, Budget files, Payroll reports and even meeting notes, underlining our commitment to being a comprehensive solution for real estate management. One distinct feature of AssetGPT is its innovative memory system. Our AI builds on past insights, preserving them in a vector database for future reference. This feature facilitates AssetGPT to evolve in tandem with its users' requirements and consistently augment the quality of its output over time.

Virtual Financial Analyst

In the rapidly progressing world of large language models (LLMs) like GPT-4, the art of crafting effective prompts is crucial for harnessing their potential. However, the dynamic nature of LLM research presents a challenge: keeping up with the continuous influx of new prompting techniques. This is where our project, the "Prompt Consultant", steps in. The Prompt Consultant aims to guide users in generating more effective prompts, not by having them chase the latest research, but by leveraging the power of the LLMs themselves. We exploit the LLM’s capacity to assimilate the best prompting resources and provide insights for improving prompts. The challenge is that LLMs, due to their static training data, are not aware of the latest prompting tricks. Our solution is to use in-context learning, incorporating the most recent prompting resources directly into the prompt. Anthropic's long-context API is a critical component of this endeavor. It's impractical to train a new model every time a new prompting method emerges, and the versatility of user queries makes common vector search methods insufficient. The long-context API allows us to include extensive relevant prompting information directly in the context. Our proof-of-concept demo uses the latest resources from learnprompting.org, embedding them in the model’s context. Users can then consult our bot, implemented on Anthropic’s Claude-v1.3-100k model, to enhance their prompts. Our preliminary results show promise, indicating LLMs' potential to stay in step with rapid advancements in their field. In essence, the Prompt Consultant bridges the gap between the rapid progression of LLM research and practical, effective usage of these models. By leveraging the LLMs themselves, we aim to make these technologies more accessible, democratizing the benefits of AI research. Our project foresees a future where anyone, regardless of their expertise, can generate high-quality outputs from these models through optimized prompting.

long long int

ThinkSync is an innovative SaaS product designed to revolutionize team collaboration through the power of generative AI. The platform integrates the capabilities of the Claude 100k context API, which is designed to enhance communication and collaboration within teams by offering AI-assisted dialogues and interactions. The problem ThinkSync is addressing is clear: current Large Language Model (LLM) chatbot solutions are tailored for 1:1 interactions, leaving a significant gap in the market for team-based applications. Teams need a way to include their LLM as an active participant, generating insights and facilitating collaboration. Furthermore, generative AI output needs to be conveniently exportable to maximize its utility in a work environment. ThinkSync's solution involves transforming a familiar team chat interface—similar to the likes of Teams or Slack—into a collaborative AI-powered workspace. Users can easily add their own API keys and initiate interactions with the AI assistant, Claude, using the "/claude" command. This brings the power of generative AI into the heart of team conversations, making it an active participant rather than a passive tool. Beyond simply chatting with Claude, ThinkSync offers the ability to invite team members into the workspace, fostering a collective interaction with the AI. This collaborative AI engagement is further enhanced by the addition of Stable Diffusion Image Generation, opening up new possibilities for creative brainstorming and problem-solving. ThinkSync aims to expand the possibilities of generative AI by exploring and integrating other generative AI APIs into a plugin "store". The platform will also offer useful commands for specific objectives such as summarizing chats and creating action items, thereby boosting team productivity. Moreover, it plans to integrate with team management and scheduling APIs, providing a seamless and holistic solution for team collaboration.

ThinkSync

Practice is extremely important for the success of therapy, but it can often be difficult to engage in on your own. Therai utilizes Antrhopic's AI chatbot, Claude, to allow individuals to practice therapy exercises in a more engaging manner. The goal is for it to work alongside a therapy regimen, improving treatment adherence and outcomes. While working with a trained professional is strongly recommended, Therai can also be used by individuals who cannot afford or easily access a therapist in order for them to gain some of the benefits of therapy. Therai.app is built using Progressive Web App technology and can be installed on mobile devices by adding the site to the home screen. After doing so, the app will have its own icon and open full-screen. User API keys and chats are stored locally on the device and remembered between sessions. The backend is an Express Node.js web server. Redis is used to allow editing the prompt and custom API keys without needing to make code changes and redeploy.

The Work

This program utilizes the latest advances in AI, specifically Large Language Models (LLMs), to deliver the best machine translation experience. Unlike traditional translation tools, LLMs are able to understand language in a more human-like manner through their neural networks. This allows them to produce translations with greater context, nuance and coherence. The program offers an intuitive user interface to leverage the power of LLMs for translating between languages. It includes features like context preservation, customizable parameters, dictionary management and interactive querying to provide greater control and insight into the translation process. By harnessing LLMs, the program is able to significantly outperform generic online translators and mimic human translators, enabling seamless cross-lingual communication.

Anthrax

Investment Apprentice (IA) is a tool that helps investment managers and analysts sort through a sea of information and analyze companies in a timely manner. The investment industry is inundated with all kinds of information, and the ability to sort through this information and sieve out noise in a timely manner is a key competitive advantage. While the industry has no issue processing structured data, processing of unstructured data is still generally done manually, requiring a high level of experience and domain knowledge. IA is a tool that can extract useful and important messages from investment articles, such as earnings transcripts, quickly and accurately. It does this by processing natural language and identifying key points, risks, and concerns by investors. More importantly, IA is impartial and objective, providing investment professionals with actionable insights to help them make better investment decisions. IA delivers results through a website. Investment professionals can use IA to quickly identify key insights from earnings transcripts, such as trends, risks, and changes from the previous quarter. IA also provides a sentiment rating for each company based on the information provided, allowing investment professionals to quickly gauge the overall sentiment around a company. IA runs in the background and is activated when new transcripts become available. Its results are stored in a database, so the process does not have to repeat when users need information. This also reduces latency and ensures that investment professionals have access to the latest insights. In summary, IA provides investment professionals with objective and actionable insights to help them make better investment decisions. Its ability to quickly process unstructured data and identify key insights gives investment professionals an edge over their competitors.

mams-nerd

Our goal is to support aspiring Youtube creators to concentrate on their objectives. Maintaining consistency and motivation often proves challenging, so we created an application where 1. users can set their goals, and the AI Coach will generate tailored go-to dashboard and an actionable to-do list. 2. users can consult with the AI Coach for any issues they might face. Problems and target: Youtube creators often face numerous challenges including maintaining a regular posting schedule, generating creative content ideas, and deciding on priorities. Solution: Our solution addresses these issues by offering a personalized, AI-powered dashboard and to do list, coupled with a conversational AI, to assist in navigating these challenges.

Tube AI

Kiku empowers company executives to excel in investor communication by anticipating the toughest questions with finesse. With Kiku's guidance, no investor or analyst query is too challenging to handle. Moreover, Kiku goes beyond mere preparation. By meticulously analyzing strategies, risk factors, strengths, and opportunities, Kiku provides invaluable insights into the company's position vis-à-vis its closest competitors. This invaluable information facilitates informed strategic decisions. But that's not all—Kiku also offers a panoramic view of the industry landscape, enabling companies to grasp the bigger picture. With Kiku by your side, conquer investor challenges, gain a competitive edge, and stay ahead of industry trends. We accomplish all of this utilizing our expert knowledge and harnessing the power of Anthropic's LLM, Claude.

LetsKiku

Our cutting-edge chatbot personality creation program utilizes ClaudeAPI large 100k context window to analyze vast amounts of data from PDFs. By carefully analyzing this data, our program is able to generate a highly sophisticated and nuanced personality for the chatbot and pack it into JSON and PNG character card formats. Thus ensuring that the chat bots are able to engage in intelligent conversations with users and provide a very customizable, dynamic and human Customer Support agent. With our program, you can be confident that your chatbot will have a personality that is sure to impress even the most discerning users.

PersonaWizard

Stori is an AI-driven game design software platform aiming to revolutionize the way video games are created. Designed to ease the process of game development, Stori helps developers, both professional and hobbyist, in building lore-consistent characters, places, events, missions, and more. The tool simplifies the intricate process of world-building in game development by leveraging artificial intelligence and vector databases. It not only reduces the time and cost associated with creating a game's world but also maintains the creative control in the hands of the developers. Its collaborative features further make it a powerful tool for large gaming companies and independent developers alike. As Stori grows, it plans to expand its capabilities, including an engagement layer and a lore marketplace, fostering a more interactive and dynamic game design community. With the potential to democratize game design, Stori aims to transform the gaming industry, making game creation more accessible, affordable, and fun.

tbd

A way to save time for users to see if a video actually contains information they might find relevant and to see when the topic in the video is first brought up. The frontend takes the video URL. The backend gets the transcript of the video (auto-generated usually by YouTube with timestamps) to Anthrophic's Claude API and get a summary using a prompt for this with timestamps for the topics. Title and description is also used to check to see if the video is clickbait or not based on the summary that's created. The result is shown on the page for the user, with whether it may be clickbait shown at the end of the summary. The presentation covers this information in more detail, as well as showing two YouTube video summaries being created in a demo at the end of the presentation video.

Team Jill

Toster, built on Anthropic's Claude, transforms user agreements into clear summaries. Better World and Anthropic share a commitment: make AI safe and beneficial. We prioritize transparency, privacy, and AI aligned with human values. We see AI as more than just a tool; it's a catalyst for progress. We are determined to break down the barriers to understanding by making the essence of user agreements transparent to all users. We stand for privacy, recognizing that every individual should be able to know and control the data they share with digital services. We believe in harnessing AI's power for the greater good. Toster is a practical application of this belief. Clarity drives understanding and awareness, empowering humans to build a better world.

Better World

Lisan is a story-based platform for language learning. In "Story" mode, the user initiates a story in a language of her choice (i.e. French) with a sentence in that language. In the feedback stage, Lisan analyzes the correctness of that sentence and provides a corrected version of that sentence. Lisan also vocalizes the sentence so that the user can check their pronunciation. In the response stage, Lisan writes her own sentence that continues the story in a coherent way, adjusted to the learning needs and current skill level of the learner. This sentence is presented in three levels of increasing granularity: (1) a vocalization of the sentence (2) a transcription of the sentence and (3) a word-by-word translation of the sentence. If the learner cannot understand the sentence when presented at an earlier stage of granularity, they are encouraged to reveal the next stage. Then, the user contributes another sentence and the pattern repeats. In "Flashcard" mode, the user can view the sentences and words used in their story by both the learner and by Lisan. They can be exported as a csv and uploaded to flashcard services like Anki or Quizlet for memorization exercises. In "Analytics" mode, the user can view (1) the full transcript of the story with all parts of speech tagged (2) a summary of the story and (3) an interactive graph visualization of a trie comprised of the words used in the story. The trie visualization is an effective way for visual learners to understand the relationships between words. For example, the words "reader", "reading", and "read" all share a common lexical prefix "read", which corresponds to a morphological root word in English.

Lisan

Revu Code is an open-source project, enabling you to communicate with your entire code base seamlessly through a chat interface. With AI-fueled insights, Revu Code empowers developers to effectively manage, enhance, and collaborate on their code, no matter their platform of choice. Revu Code is designed as three interconnected components: 1. **API**: A powerful and intelligent back-end that processes, analyzes, and manages your code base to provide relevant information and recommendations. 2. **Flutter App**: A versatile and responsive front-end built with Flutter, allowing you to access Revu Code as: - **VSCode Extension**: Integrated directly into your favorite code editor. - **Web App**: Accessible from any browser for ultimate convenience. - **Mobile App**: Stay connected to your code base on-the-go. - **Desktop App**: Enjoy the full potential of Revu Code on standalone desktop applications for Windows, macOS, or Linux. ## Experience the transformative benefits of Revu Code: - 🔎 **Ask Complex Questions**: Delve into your code base with Revu Code's chat interface, answering challenging questions and deciphering relationships with the power of AI. - 🚀 **Accelerate Answers**: Receive instant, relevant insights by typing your queries directly into the chat interface, whether in the VSCode Extension, Web App, Mobile or Desktop App. - 📈 **Improve Code Quality**: Enhance performance and readability with intelligent refactoring and optimization recommendations across all your platforms. - 🔧 **Identify & Fix Issues**: Automatically locate and rectify inefficiencies, security vulnerabilities, and deprecated dependencies for polished code. - 💬 **Streamline Collaboration**: Facilitate effective code reviews and seamless team communication with in-context, chat-based collaboration. - 🧠 **Contextual Documentation**: Access code examples and contextual documentation instantly, across devices, to boost your productivity.

ConceptaDevs

Imagine reading a complex text, say Keynes's General Theory, and struggling to comprehend the intricate concepts. Now, imagine an intelligent assistant that takes these abstract concepts and links them to real-world scenarios and current events you care about - making learning not only simpler, but also engaging and relatable. That's exactly what MintAI offers! MintAI is your AI-powered personal learning assistant, transforming the way we approach learning complex material. It bridges the gap between theory and practice, enhancing your understanding and retention of information. But, it's not just about simplifying the material - MintAI is built on a sophisticated AI that understands your unique learning style and personalizes the content accordingly. What's more? You are part of an active learning community! MintAI encourages you to upload materials, interact with the AI, and comment on real-world events with an understanding of theoretical concepts. As a result, you're not just learning - you're applying your knowledge and contributing to the learning of others. At the core of our business model is a subscription service that offers access to premium features and personalized content. This ensures a sustainable and continuous platform enhancement, making MintAI your lifelong learning partner. Join us as we revolutionize the learning experience, making it easier, more engaging, and highly personalized. Welcome to MintAI – your future of learning.

SparkO2

Introducing a game-changing, AI-powered trading app designed for the contemporary investor. Whether you're looking to build a tailored stock portfolio or explore AI-led strategies, our platform provides the perfect solution. Explore these cutting-edge features: - **ChatGPT-Driven Strategy:** Effortlessly buy and sell stocks through a sophisticated AI strategy powered by ChatGPT. Trading is now intuitive like never before. - **Sentiment-Based Trading:** This unique feature bases stock purchases on the tone of news headlines, adding an extra dimension to your trading strategy. - **Community Sourced Strategies:** Tap into collective wisdom with proven trading strategies imported directly from established financial communities. - **Replicate Winning Strategies:** Mirror the strategies employed by prosperous funds and top-tier investors to grow your wealth. - **Performance Comparison:** Measure the success of your strategies against major indexes, successful funds, or each other. Our platform, already integrated with Alpaca, is looking forward to partnerships with Degiro and more. We're also expanding into crypto trading with integrations like Binance, Coinbase, and Kraken on the horizon. Initially launching in US markets, we aim to expand into other stock and crypto markets for a global trading experience. As a tech-first, socially responsible company, our platform resonates with digital natives familiar with Robinhood, Uber, and Netflix. Committed to promoting socially responsible investments, we pledge to support environmental and social projects. Embrace the future of trading. Your wealth, your way.

AI traders

Before the main discussion, let's look at the long-term memory mechanism of the human brain that we imitated. We are not famous neuroscientists, so to put it briefly, information created from various information is stored in the hippocampus as 'short-term memory', and later some of this memory is stored in the cerebral cortex as 'long-term memory'. Multiple pieces of information are combined into new information in the hippocampus, stored as short-term memory. Later, parts of this short-term memory are transferred to the cerebral cortex, where they are stored as long-term memories. We borrowed part of this method for AI and designed this function (extension). In this method, there are two AI's: an AI (hereinafter main AI) that conducts a conversation with the user, and an AI (hereinafter auxiliary AI) used to remember the conversation. When a conversation between the bot and the user takes place once, the exchange between the bot and the user in one conversation is stored separately in a file. At this time, the auxiliary AI analyzes the stored conversation file and stores several topic clusters of the conversation together. Later, when the auxiliary AI analyzes the user's speech in the conversation, if there is a conversation with topics similar to the current speech in the previous conversation, the main AI is also referred to the corresponding conversation (past conversation with similar topics). Through this method, semi-permanent memory is stored according to local storage space.

Turbulence

Welcome to the future of online shopping - AIGC, the Advanced Instagram Global Community for Shopping. Immerse yourself in a world where cutting-edge technology and retail therapy converge. Our community is teeming with dynamic virtual AI influencers, each possessing a unique style and personality, and each powered by the trailblazing Anthropic AI. These AI connoisseurs tirelessly scour the globe, unveiling the very best products from every corner, and presenting them to you through engaging personalized posts. With AIGC, the world of shopping is literally at your fingertips. Follow your favorite AI influencers, engage in illuminating conversations, or even explore your creativity by designing your own virtual influencers. Whether you're a seasoned shopper or just starting out, AIGC offers a versatile platform where you can discover and explore new products in an entertaining and immersive way. As part of AIGC, every day is a shopping adventure. Join us as we revolutionize the way you shop and transform the shopping experience. Enjoy the thrill of global shopping exploration from the comfort of your home, and be part of a community that’s at the forefront of the AI-driven shopping revolution. You are not just shopping, you are shaping the future of retail. Welcome to AIGC.

Team searchSmart

LLM naturally can work as a flexible classifier: it's so flexible that you don't need to specify the format of your input or clearly explain your classes. This project uses this trait of LLMs to solve a problem many people face in life: a messy Desktop or Downloads folder. It might contain a lot of screenshots, bills, memes from Internet, legal forms your company or school asked you to fill, software, etc. We want to put them into the right place, so that I can find them easily when I need them, but they are most likely named like "241223.pdf" or "IMG_1231.jpeg". This project read the filename, the file type and the content, and allows user to specify which folders they want the files to be organized into. Then, AI will take care of the rest.

Organizer

Screenplay Similarity Search will be part of a much larger AI-driven screenplay analysis platform, which helps screenwriters improve their work through detailed analysis of their themes, characters, objectives, structure, writing style, and more. We built the broader product with GPT-4 and launched it at ScreenplayIQ.com (which already has paying customers), but for this hackathon we made a new product simply to find matching screenplays for writers to analyze together. Thew new product we built for the hackathon, Screenplay Similarity Search, creates value for writers by pointing them toward other work which matches their tone and subject matter. It will be even more impactful when it is hopefully integrated into ScreenplayIQ, so writers will be able to compare much deeper analysis (which we plan to use Anthropic for as well, replacing GPT-4 for most of the tasks, since we've seen Anthropic's LLM be far more capable than GPT-4 in certain kinds of analysis). So search is just a small part of helping screenwriters improve through AI-driven feedback and comparison.

Screenplay Search

Prereader is an innovative application developed to help students familiarize themselves with complex texts before reading them. The app uses the extensive contextual understanding of an AI named Claude to extract key and methodological concepts from academic materials. By transforming text into audio, using the googleTextToSpeech python package, students can listen to these ideas before delving into the text, supporting their comprehension. Prereader is designed to deliver information in an unpressurized way, with no need for memorization, reducing the stress of encountering new academic content. A unique feature of Prereader is the provision of an offline audio file. This allows students to utilize the app even without data access, whether they are commuting on public transportation or enjoying a quiet walk outdoors. The application combines the advanced capabilities of Claude, the convenience of the googleTextToSpeech package, and the streamlined efficiency of Streamlit for deployment. In summary, Prereader is a powerful tool aimed at improving comprehension and information retention for students tackling challenging academic materials.

Finsbury Features

Ask Arivx is a revolutionary AI-powered virtual assistant designed to augment and accelerate knowledge discovery. It harnesses the potential of the latest advancements in machine learning and natural language processing to deliver a highly interactive, intuitive, and context-aware question-answering platform. Unlike standard search engines or digital assistants, Ask Arivx engages users in meaningful dialogue, interpreting inquiries in a nuanced way to provide the most accurate and comprehensive answers possible. By simulating a human-like conversation, Ask Arivx encourages active learning and promotes a deeper understanding of various subjects. The platform serves as an essential tool for a wide range of applications - from academic research and learning support to business insights and personalized daily assistance. Furthermore, its flexible, scalable design ensures seamless integration across different platforms and devices. With Ask Arivx, we are not just redefining the way users search for information; we are building a future where technology empowers individuals to harness the power of information more effectively and efficiently.

Ask Arivx

We often participate in tens of, if not hundreds of, group chat via IM apps (discord, wechat, telegram, ...) and are not able to follow all the topics being discussed. It would be of great help if there is a tool that can summarize it for us, once a day. ChatSummary bot is a tool that can be added into any group chats and records messages continuously. At the end of each day, it will generate a report of all the hot topics today, along with the time it happened, the summary of the topic, and noteworthy contents in this topic. This helps you to quickly and efficiently gather information from your social circle.

CopperFrenz

plateful: It's a common sentiment (one I personally share, hence this idea!) that recipe and meal planning apps today are too complicated for the average home cook, especially with the limited time we have in the hustle and bustle of everyday life. They zap the joy out of cooking with tedious data entry, convoluted recipe organization, and confusing interfaces. For this hackathon, we have decided to build plateful, a delightfully simple AI cooking companion. The aim of plateful is to make home cooking straightforward, effortless and accessible. Unlike other apps that require slogging through calculations and management, our solution gets you cooking in seconds. plateful streamlines the planning process by automatically generating ingredient lists - even from the most vague of inputs! - and providing one-tap meal options tailored to the user's tastes, whether they're cooking for themselves or a whole band of picky eaters. plateful will handle the hassle so users can fall in love with cooking again. Let's let Claude's talent as a chef shine through.

Team Helix

Claude Copilot for CLI, also known as C3, is a user-friendly command line tool designed to integrate smoothly with your terminal. Powered by the Claude AI Assistant, C3 makes it easy to generate commands using natural language prompts with the `c3?` prefix. C3 is inspired by GitHub's Copilot for CLI tool. For example, you can enter `c3? list all .js files in my directory`, and C3 will provide `ls *.js`. Need to include subdirectories? Simply use the follow-up option, clarify your requirement, and C3 will adjust its output to `ls **/*.js`. Additionally, C3 enhances the command line experience by offering clear explanations of various flags and parameters for less familiar tools. To see a quick demonstration, please visit the following link: https://asciinema.org/a/fBjlXuDf6qGOdPprbzqNoJHq3.

Team Marvin

Workforce Continuous Intelligence is a no-code UI that connects to external resources. - Tasks can be either simple tasks or chats. - Workers can be either AI or Humans. - Prompts can be generated from a combination of external resources, and are dynamically generated based on the content of those resources. - Resources are updated in real time, and triggers can be set up to execute a task in response to a change in a resource. - Resources can also be pushed to in order to take action on behalf of a task. For example, the Github Resource will store the output of at ask into a file in Github. - Trackers can be integrated. Tasks will pull tickets from Trackers as soon as they show up in the Tracker, and they can also generate new tickets in a tracker. - Transforms can be used to extract data from the output of a task and segment it out into individual pieces.

DevBots

AI technology can be used to model the use of nanoplatinum in antimicrobial applications. By analyzing the physical and chemical properties of nanoplatinum and its mechanisms of interaction with bacteria and human cells, the effectiveness of nanoplatinum in the treatment of bacterial infections can be predicted. Also, AI technology can use big data analysis to explore the optimal drug delivery of nanoplatinum and optimize its application route and dose to improve its efficacy and reduce adverse effects in patients. In addition, AI technology can be used to monitor and predict the drug metabolism and side effects of nanoplatinum to develop safer and more effective treatment plans. Through these applications, AI technology can greatly enhance the value of nanoplatinum in the field of antimicrobial applications.

Nano platinum antibacterial

Start & Grow your Business Effortlessly via Speech Agent: It’s hard to be alone, especially when running or trying to start a small business. Our idea is to create a voice-interactive assistant aimed at supporting small business owners. This app will fully handle a variety of tasks, such as budget optimization, tax filing and employee documents, with the owner providing final review over the results. Through repeated interactions, the assistant will develop predetermined workflows based on user patterns, analyze to deliver actionable insights, flag areas of concern, and even discuss new business propositions.

Noisebridge AI

I am Taiwanese and not good at English, so my English might sound strange. I apologize. Claude Story Writer Webui allows writers to input inspiration for stories, and it generates a world view, characters, allies, enemies, and conspiracies based on that. It provides prompts to assist writers in conceptualizing their stories from various perspectives. These concepts can help writers unleash their creativity because they don't have to meticulously fill in the details of the world view themselves. Depending on the user's input, the script generated by Claude can provide users with great ideas, although it may sometimes be unstable. That's why I have been trying to update and use better prompts. The script is saved as an "abbreviated_story.json" file, and if the user finds the script generated this time to be good, they can keep it. My next attempt is to turn this script into an interactive game, where Claude reads the script and transforms into a communicative role-playing game, similar to tabletop RPGs. Users can freely act within the scenes explained by Claude using natural language. However, I found that Claude tends to deviate from the original script after a few rounds. Therefore, every five user responses, Claude conducts a review of the current story and script, examining the current situation and aligning it with the script to analyze and arrange for the new story. Users can explore within the script generated by Claude, or they can input their own script and have it generate scenes and interact with them.”

Ayaori

Traditionally we are used to the process of data research with jupyter, whether it's notebook, lab, or notebook intergrations in IDEs. ClaudeCode is an assistant connected into your jupyter session and code for you. Like your reseach pair, it can take over the data you left, listen to your requirements and generate code then execute it for you. Whether it's a modification of the data or train a model. It will record history conversations and code generated. So it will keep context. And it's also aware of the data currently in the memory, especially for pandas dataframes' columns, meanings and dtypes.

FractalParrot

Ask pointed questions about a given playlist and get back a summary, key points, and related timestamps generated via AI! 🤖 Could be podcast series, a learning series, or something completely different! Can take in even very large/long series (tested on ~150 ~2-hour long podcasts)!Ask pointed questions about a given playlist and get back a summary, key points, and related timestamps generated via AI! 🤖 Could be a podcast series, a learning series, or something completely different! Can take in even very large/long series (tested on ~150 ~2-hour long podcasts)! This tool can take a YouTube transcript from one or more videos to be used to answer questions on a topic. The output will include a generated overall summary and generated key points from the video(s) by reading select parts of the transcript. The output will also include links to the relevant video, timestamped to the specific quote/snippet related to its respective key point. This tool can be useful to learners going through a video series playlist to review or identify where the series talks about a topic. It can also be used for educators in creating lessons from a series of videos. It also can be used for more casual enjoyment such as reviewing what the hosts have said on a particular topic. This use case is especially relevant for podcasts where hosts may revisit the same topic across multiple topics. Although Anthropic's Claude model can take in 100k tokens, this still creates a limit to what's read in by the LLM. This project will attempt to read in all the selected transcripts for the available model but if the transcript is too big for even the beefiest model, the tool will strategically select portions of the relevant transcripts based on the user fed question.

Vector

Thanks to AI, it can: - Show users the way - Claude will read the list of channel names and descriptions and respond to user with the channel that is the most appropriate for their query. - Analyse logs - The bot can download recent messages from a user and analyse them for breaches of guidelines or unprofessional behaviour. It can also analyse channel history and give a rundown of important information. - Decision-making - When analysing user messages, Claude considers the severity of guideline breaches and suggest a course of action. This will help admins decide if a ban/kick is justified. - General user support - Claude can answer discord-related questions, with the option of asking followup questions (within a limit). - Message translation - Claude can translate messages to multiple languages.

Katworks

NextCoding is an AI project that generates code based on project descriptions and instructions. It simplifies the coding process by leveraging artificial intelligence to interpret and transform human inputs into functional code. This innovative project eliminates the need for manual coding, saving developers time and effort. With NextCoding, developers of all skill levels can automate repetitive tasks and focus on higher-level problem-solving. By continually learning and adapting to evolving coding practices, NextCoding ensures the generated code is reliable, optimized, and aligned with industry standards. It has the potential to revolutionize software development and increase productivity.

kylin1020

The concept hinges on the use of the Language to Language Model (LLM) command to generate prompts for the Stable Diffusion (SD) process, which then creates images. The LLM command converts linguistic instructions into AI-compatible formats, guiding the SD process. SD, a common AI technique, spreads information through a system until a stable state is achieved, making it effective for image generation. The LLM command creates prompts that direct the SD process to create specific images. These prompts, which are linguistic instructions, are designed and coded strategically to guide the image generation process towards a desired outcome, like an image of a sunset over a calm lake. Once the SD process starts with these prompts, it gradually transforms the initial state (a noise pattern or blank canvas) into a complex image that aligns with the prompts. This diffusion process ensures the image evolves in a stable, controlled manner, closely following the instructions in the prompts. The outcome is an image that accurately reflects the original LLM instructions. This isn't the end, however. The image can then be used as input for further LLM commands, which create 'img2img' prompts for another round of SD, further transforming the image based on new instructions. This enables the creation of a variety of images, each interpreting the LLM-generated prompts uniquely. This approach, combining the linguistic versatility of LLM commands with SD's transformation capabilities, creates a cycle of image generation and transformation. This cycle opens up extensive possibilities for creative expression and AI-assisted design, with each cycle potentially birthing a new artistic creation.

Geek1024

I propose an open plugin architecture that allows language models to access external data and capabilities through plugins. Language models today are limited to their training data, but real-world AI requires access to vast, disparate data sources. An open plugin architecture addresses this by enabling plugins for search and more. The architecture defines an interface for creating and installing plugins. Developers can then create plugins accessing proprietary data, open data, and online services. The plugins run externally but programmatically interact with the language model, feeding data and retrieving responses. This idea came from the need to extend language models beyond their training data. The open architecture means the capabilities of models can grow over time through the development of new plugins. This has the potential to make language models much more useful and scalable.

AI enhancement plan

ThinkATive offers several key features: 1. Confidential and Non-Judgmental Environment: The AI counseling platform creates a safe space where individuals can openly express their thoughts, feelings, and concerns without fear of judgment or stigma. 2. Personalized Counseling Sessions: Through advanced algorithms and machine learning, the AI system can analyze users' responses and provide tailored counseling sessions based on their unique needs and circumstances. 3. 24/7 Availability: Recognizing that mental health issues can arise at any time, the AI service is accessible round-the-clock, allowing individuals to seek help whenever they need it, even outside traditional working hours. 4. Multilingual Support: The AI platform is equipped with language capabilities that cater to diverse linguistic backgrounds, ensuring that individuals from different communities can access the service in their preferred language. 5. Resource Recommendations: The AI system can suggest relevant resources, such as articles, videos, or self-help materials, to supplement counseling sessions and empower individuals to continue their personal growth and healing. 6. Referral Assistance: In cases where additional support is necessary, the AI service can provide referrals to local mental health professionals or community organizations, ensuring individuals receive appropriate and specialized care. By combining the convenience and scalability of AI technology with empathetic and compassionate support, this AI-powered psychological counseling service strives to bridge the gap in mental health care for marginalized populations, promoting well-being, resilience, and inclusivity.

Arcalive

This AI-driven system is designed to ensure compliance within the context of email communication by adhering to internal policies and external regulations. It scans and analyzes email content in real-time, identifying potential compliance issues and alerting the relevant parties in case of deviations. A unique feature of this system is its access to a repository of compliance documents, enabling it to stay current with the latest rules and regulations, thus ensuring its compliance checks align with the most recent requirements. The 100K token window plays a crucial role in the operation of this system. Given the complexity of compliance requirements and the diverse range of topics that can arise in email communications, a large context window is necessary to fully understand and analyze the content of each email. The 100K token window allows the AI to consider a wide span of context when determining compliance, including previous email threads, referenced documents, and the specific wording of regulations in the compliance documents. This large context window thus ensures a higher degree of accuracy and comprehensiveness in the system's compliance checks.

aitrailblazer

Personalized guidance. The AI is designed to have a "persona" of an experienced life coach and provide tailored responses and questions to guide the user through a coaching session based on their unique goals and needs. Achieve It Coach provides the following to its clients: Structure and accountability. The AI follows the well-known GROW model for coaching, which helps provide a structured framework to keep the user progressing and accountable to the goals and actions they define. New perspectives. An AI coach can provide users with an outside, objective perspective, which can help them gain new insights into their situation, obstacles, and solutions. The open-ended questions are aimed at facilitating the user's self-discovery and problem-solving. Convenience and availability. An AI life coach would be available on-demand whenever the user wants guidance or support. They do not have to schedule time with a human coach and can have shorter, more frequent interactions. Privacy and anonymity. Some users may feel more comfortable opening up to an AI than a human coach. Interacting with an AI can provide privacy and anonymity for those discussing sensitive life issues and challenges. Continuous progress. The AI is designed to provide follow-up and continuity between conversations to keep the user advancing, even with short interactions. It can recall the context from previous discussions to build on momentum and progress.

Achieve It Coach

Struggling to stay abreast with the dynamic advancements in the field of AI, often overwhelmed by the sheer volume of incoming information? Let your worries be a thing of the past with researchAssistant - your personal AI-powered ally. Simply feed it a list of URLs, and it will dutifully scan, summarize, and rank the information, presenting you with a detailed analysis for each source. The beauty lies in the seamless integration of these analyses into Obsidian, making the once overwhelming information instantly navigable. ResearchAssistant transforms your digital research experience, ensuring you never miss a beat in the fast-paced world of AI (or any other field of your interest).

KnowledgeNavigator

Our project tackles the urgent challenge of climate change and emphasizes the need for collective efforts. With busy schedules, it can be challenging for individuals, especially students, to find time to research the environmental impact of their actions. Misconceptions and limited knowledge can lead to unintended harm, such as incorrect recycling practices that can be worse than not recycling at all. To address this, we have developed a personalized sustainability tool. This innovative app offers fast and user-friendly assistance to help users make informed choices. It provides valuable insights into the environmental impact, feasibility, and effectiveness of various practices. Additionally, the app suggests improvements and provides additional information to guide users towards more sustainable behaviors. For example, when considering the disposal of styrofoam boxes, the app evaluates the trade-off between washing them for reuse and throwing them away, taking into account factors like water consumption and waste generation. By considering these variables, users can make more informed decisions aligned with their sustainability goals. The app goes beyond generic advice by explaining the reasoning behind recommendations, fostering understanding, and allowing users to ask questions to enhance their knowledge. It takes into consideration specific factors such as location, age group, and workload to offer tailored advice. This context-specific approach ensures that the recommendations are relevant and applicable to the users' unique circumstances.

Sustainable Initiatives

Telegram Chat Summarize is a Python application that uses the Telegram Bot API to access the chat history of a Telegram group and generate daily summarizations. The application is built using Python 3.x and pip for package management. The summarization process is powered by either the Claude API or the OpenAI API. The application requires several environment variables to be set before it can be run. These include the API keys for the chosen summarization service, the API token for the Telegram Bot API, the API ID and hash for the Telegram app, and the Telegram channel ID for the developer to receive messages. The application is designed to be run on a local machine or server, and can be started by running python main.py in the command line. Once the application is running, the user can interact with the bot using various commands, such as /start, /set_chat_name, /summary, etc. Overall, Telegram Chat Summarize is a powerful tool for users who want to stay up-to-date with their Telegram group chats without having to read through every message. By using advanced summarization techniques, the application is able to provide concise and informative daily summaries that can be accessed easily through the Telegram Bot API.

solo developer

Our project utilizes the Anthropic Claude to understand long, complex texts such as HOA disclosures or inspection reports within the real estate industry. Leveraging the model's capability to understand and generate human-like text, we aim to offer the user a quick summarization of the lengthy documents. The model will pinpoint and highlight the key points and risks in the document. Furthermore, based on user requirements and inputs, our AI assistant will provide a recommendation score. This score will aid users in making a more informed decision about the property they're dealing with. To ensure that our solution meets user needs effectively, we plan to design it to conduct constructive conversations based on further user questions. This conversational aspect will provide users with deeper clarity on specific points, thereby fostering better understanding and decision-making.

Soul AI

AI Interview Assistant is an artificial intelligence platform that automates the interview process. Candidates provide initial criteria like role and experience. The system then generates a customized interview with multiple follow-up questions for each answer. Through progressive Q&A, AI Interview Assistant assesses candidates thoroughly. After the final round of questions, the system produces an interview analysis and suggestions for candidates to improve their performance. User feedback also helps the AI learn and optimize its interview skills over time. AI Interview Assistant saves time and resources by handling the initial screening and shortlisting of applicants. The personalized nature of the interviews also leads to a better experience and more accurate match between candidates and opportunities. With its ability to understand language and evaluate qualities beyond basic qualifications, AI Interview Assistant aims to transform the hiring process through enhancements in both efficiency and efficacy.

KIKIYO

We create an interactive role-playing novel game. You will play the game by entering statements. The assistant played by Claude will respond you information from script. You will find clues and submit your reasoning at last. We are a team dedicated to creating immersive and interactive gaming experiences. Our game aims to offer a rich interactive experience, and we believe that the Claude-100K model API can be a valuable tool in achieving this goal. We plan to use the model API to generate dynamic and engaging text-based interactions and narratives. This includes but is not limited to: Character Interactions: Utilizing the model API, we aim to create non-player characters (NPCs) that can carry out complex and engaging conversations with the player. These interactions will be context-aware and responsive to the player's inputs. The first challenge is that Claude may respond with hallucination. We designed the prompt to weaken the hallucination. The second challenge is how to force Claude not tell the player the truth directly. We know, Claude is a very safe AI. So, it may not tell a lie even in a role-playing game. We designed the information input to make sure Claude won't Reveal the follow-up plot. We will take you more immersive game experience in the future!

FlyingKittens

The Professional Translator is an advanced language model-driven system designed to provide accurate and in-depth translations with a deep understanding of professional knowledge. By leveraging the in-context learning capabilities of a large language model and employing semantic search techniques for knowledge acquisition, this project aims to revolutionize translation services for technical and specialized domains. Key Features: 1. Contextual Understanding: Our translator will utilize a large language model to analyze the input text in a specific professional domain and generate translations that are contextually accurate. By considering the broader context of the text, including industry-specific jargon and concepts, the translations will be more precise and cater to professionals' needs. 2. Semantic Search for Knowledge Acquisition: Implement a semantic search module that can crawl and extract relevant information from authoritative sources, research papers, industry forums, and other trusted platforms. This module will continuously expand the knowledge base and improve the system's ability to handle specialized terms and concepts.

Vanguard

The fundamental act of reading is essential in today's knowledge society. Reading is not just about decoding words. Reading requires understanding vocabulary in context, analyzing structure, inferring meaning, and synthesizing information—complex skills that many struggle with. The time is now to eliminate the gap between the critical importance of reading comprehension in our society and the challenges that individuals encounter in their reading journeys. Goal: Deliver a thoughtful, AI-empowered approach to significantly enhance how billions of people read, learn, and access information. In May 2023, Anthropic expanded Claude’s context window from 9K to 100K tokens. There are many opportunities for using this capability! Let’s start with reading! "Reading Boost," is an AI-empowered process that individualizes reading comprehension for students, educators, professionals, and individuals with reading or learning difficulties. The process has four steps – plan, implement, review, and improve. AI’s role is to augment each step, maintain metacognitive knowledge, and produce counterfactual strategies to improve reading tasks.

Knowledge Harvesting

Congruity.AI provides a diverse set of tools and options for business users and consumers alike to interrogate the consistency of content provided by large language models. It does this by first returning a varied sample of responses, utilizing different API parameters as well as prompt variation. We then ingest and analyze the results to present visualizations and other tools for users to investigate the quality and consistency of generated text. These approaches include an ngram census (portrayed on the text), similarity matrices (showing how similar responses are to one another), links to external sources for verification, and a comparison to constitutional rules or values for the language model. Using these complementary tools, users can drill down to identify elements of generated text that are consistent with the facts, with one another, and with corporate or personal values.

Congruity

Claude is an AI assistant developed by Anthropic that can understand natural language and generate coherent responses in a conversational context. It is trained on a massive dataset using self-supervised learning which allows it to develop a broad range of knowledge and capabilities. Rimworld is a popular sci-fi colony simulation game where players control characters in a futuristic space western setting. The game generates stories and events that shape the player's experience in an unscripted manner. This type of emergent narrative is engaging but can sometimes feel repetitive after many hours of gameplay. A storytelling modification for Rimworld aims to enhance the narrative experience using AI such as Claude. The mod can generate more varied and complex story events by understanding the current game state and inventing new scenarios that fit within the established lore and tone of the game world. For example, instead of generic events like "battery explosion" or "colony marriage", the AI may create multi-part storylines following specific characters, introduce compelling villains and plot twists, or build up and resolve long-running mysteries. The integration works by feeding details about the player's colony, characters, tech progression, and past events what had regular storyteller into Claude. The AI then responds with suggestions for new story beats that can be translated into in-game events, quests, items descriptions, and character interactions. Instead of pulling from a fixed set of possibilities, each new game experience can feature completely unique stories shaped by the players' actions and the AI's limitless creativity. With the help of natural language generation, Rimworld can achieve new heights of unscripted storytelling and truly feel like every player is experiencing their own science fiction epic.

RefernoMe

I work educational technology, a problem we have currently is a influx of student devices being broken (we're not a one-to-one school, all devices are shared) without the knowledge of which school is causing the most damage in cost. With access to Claude, I'd like to implement a program with Claude to address this problem. I'd do this likely using excel/spreadsheet and a RFI board-tag tracking system. I've not brought this up to my boss in fear Claude may not be capable of such a feat- but I'm hoping with access and a bit of brain power, I can manage to get it to work. With Claude's assistance, my primary goal is to exchange the more expensive devices of trouble schools for lesser devices; Such as Lenovo Thinkpad 15's (L15s) that cost upward to ~$1000 a piece- I wish to find which school (and maybe classrooms) break the most of these and replace said devices with much cheaper alternatives. Chromebooks for example cost around ~$100, so in turn of one broken L15, it'd take 10 broken chromebooks until the price would match a broken L15s. My idea is to feed Claude a ton of information that we're in the process of getting and making Claude output more eligible and compact information. Maybe even use Claude to provide solutions with trouble schools/classrooms if they're capable of doing so.

Info-Tech Technician Moose

This submission is an AI agent that can conduct realistic and immersive job interviews. This agent allows the user to repeatedly practice the interview and to get used to the sorts of questions that would be asked for that specific type of job. The agent will provide feedback on the answers and the overall performance of the interview once the interview is complete. It is designed using a basic chat interface for the hackathon but the team plans to add a text to speech interface for two way oral interviews and a virtual reality mode for deeper immersion. Interviewing skills are not often underdeveloped by the general population, and this project aims to allow users to greatly refine those skills.

Humans are Resources

Chronic back pain is a major health crisis with significant impacts on individuals and economies. In the US, managing and treating back pain costs more than cancer and heart disease combined. Around 80% of adults experience back pain at some point in their lives, highlighting its widespread reach. Current solutions for back pain often involve invasive procedures or provide temporary relief without addressing the root cause. To address this, we're developing an innovative platform based on 'The Mindbody Prescription'. Our platform focuses on mindbody medicine and offers an interactive course to help users understand the connection between their mind and body, empowering them to alleviate their pain. Our platform also includes an AI coach that guides users through their healing journey, providing personalized advice and support. This AI assistance mimics human interaction, offering real-time feedback. We recognize the importance of community support, so our platform incorporates a community space where users can share stories, learn from others, and build a network of support. By merging health education and technology, our platform represents a significant advancement in chronic pain management. Our vision is to make mindbody approaches accessible to everyone, transforming how we perceive and manage chronic pain. Our goal is to empower individuals to take control of their health and well-being, improving their lives.

feepfoop

A web application to help people learn new words in a new language. We will use Claude to generate storylines and then put those storylines into Stable Diffusion to generate comic strips. We use Chain of thoughts to divide the task (The quality of the story relies on Claude LLM ’s creativity): Thought 1: I give you a word, you generate a vivid story, and the main idea of the story is to explain the word so that the person who has seen it can make a connection with the word in his or her head and remember the word by understanding it. Thought 2: please break this story into smaller pieces, each piece should not exceed 60 words Thought 3: generate a detailed stable diffusion prompt for every part of story Then, we send these prompts to stable diffusion backend service to generate images.

Chatty Tutor

Our project, Enigma, is an innovative and captivating game that brings together the world of situation puzzles and artificial intelligence. By leveraging the power of Claude, a generative AI, Enigma offers players a unique and immersive gaming experience. Participants are challenged to uncover the truth behind complex scenarios by asking clever and strategic questions, receiving only "yes" or "no" responses. Additionally, Enigma utilizes stable diffusion to generate visually stunning images as puzzle prompts or hints, providing players with valuable visual cues to unravel the mysteries. With Enigma, players will embark on a thrilling journey of deduction, ingenuity, and problem-solving.

LangGame

I am a brazilian aspiring writer and for some time now I have been exploring AI LLM to assist me in my writing related endeavors. AI can be a very useful tool for fully extracting the potential of your creativity, allowing for very in-depth world and character building. Although tested with other AI's, Claude seems to be exceptionaly good at this, both the prosaic and poetic aspects far exceeding those of the competition. I believe access to Claude would be very benefitial to my project and the Hackathon itself a great opportunity to get a step closer to my goals of becoming a sucessful writer some day.

Mnemosyne

Language Tutor allows students of foreign languages use research-backed techniques in second language acquisition, allowing them to receive target language practice at their current still level. The problem with language learning today is that it's difficult to find input that's comprehensible to the student. The language tutor has 2 main features to address this problem. The first feature is a conversational chatbot that speaks your target language at your level, providing you with gentle guidance on grammar and vocabulary mistakes along the way. The second feature is a story-generating tool that allows you to access graded readers at the reader's level.

Ollie

This basic userscript utilizes Claude AI to provide Twitter users with assessments of tweet sensitivity. It checks the text of a selected tweet to assess how considerately the idea is expressed against criteria given by the user. The assessment includes a short explanation of why the tweet may be insensitive. The goal is to allow users, specifically non-native English speakers, to learn harmless tweeting through real-world examples. By seeing how others conveyed ideas in both sensitive and insensitive ways, people can better understand the nuances of phrasing and avoid toxicity. This approach is more effective than simple translation because it promotes learning. Currently a manually-operated proof of concept, the project could become an automated browser extension. For now, users provide their own API key, but in the future it can be monetized, offering access on a low-cost subscription basis. As tweets are short and context-limited, AI analysis should remain affordable. Some points of note: To try it out you'll have to install a userscript managing browser extension such as Violentmonkey or Tampermonkey. This script it designed to be used specifically with open-source Twitter front-end called Nitter, the official instance of which is available at https://nitter.net . And, of course, you'll have to use your own Anthropic Claude API key.

Function Junction

LinkMotionTalk is an innovative game-model expression-driven system that offers customizable and user-friendly programs for semantic sentiment analysis system. The system utilize the power of Claude's 100K large language model of contextual semantic search to provide an advanced solution for game dialogue production and scripting. Additionally, the technology can be extended to TRPG players and virtual hosts through LiveMotionTalk, thereby offering a seamless and immersive experience for users. Overall, the project aims to revolutionize the challenges surrounding game dialogue production and bring a new level of sophistication to the gaming industry.

Tresford AI Lab

All of the currently available AI-assisted writing tools (that I’ve seen, at least) are based upon the goal of producing finished prose. This is fantastic, and is incredibly useful for small sections of text, but all of their systems tend to fall apart when they are applied to longer pieces. Writing a novel is different from writing an email. As an alternative, I propose a more layered system of content generation, with the AI acting as a writing assistant along the way. This system must allow for the following: -Rapid creation of new content, with the ability to overcome “writer’s block” -The ability to generate tens of thousands of words without entering a hallucinatory spiral. -Complete customizability, since the needs of each writer will be unique -Adjustable “autopilot”, to retain the feeling of actually writing a novel I believe that I have succeeded at a prototype of such a system. The UI isn't there yet, but the tools WORK.

The USS Galaxy

Very often couples can't discuss their problems with each other because they are emotionally involved, and the discussion usually turns into personal fight without any reasoning, or empathy. They are not even able - and after a while not even try - to understand each-others point of view, while that would be essential to solve a problem. Our project is to give these couple a solution where they can express their problem, their views to the CalmClash app, and the app simulate a constructive and emphatic discussion between the two, keeping mind their priorities and views, but do it with empathic and a constructive way. Reading that simulated discussion they can get a picture what the other party would say if (s)he would have the chance to have a calm discussion. It's like a pair therapy assistant which provide a secure and wise channel that can provide a calm, constructive discussion in an emotionally overheated situation.

CalmClash