Project Info
Inspiration
Speakeasy was inspired by a desire to better connect and empathize with our friends in both in person and online interactions. Often, we share a large portion of our lives online, but don't have time to stay updated with our friends. We were looking for a way to be able to talk to our friends without missing any critical updates to their situation.
What it does
To stay in touch with our friends and family, our users will run our application on their desktop device, and later through wearable electronics. With the consent of both parties, we source public data from their social networks such as X.com (formerly Twitter) and collect essential information they have shared with the internet. This includes light hearted posts like puppy pictures, milestones like graduation, but also serious events like a death in the family. Our goal is to support our users through conversation guidance and empathetic warnings. In this way our social network can help our users connect in a closer and meaningful way when spending time with each other, online and in person.
How we built it
We built our demonstration application with a front end using Cursor for a palatable interface with functioning interactions and room for scaling with HTML and CSS. In our back end, we divided the work, from AI integration, web scraping, and natural language processing. This ranges from understanding the images and messages in different tweets to sending sending the audible AI message through a simulated voice. We used PyTorch, LMNT, playwright, and JavaScript. Additionally, we also planned for the future with integrating wearable electronics, where we have a homemade heartbeat sensor to connect to our AI and share when a conversation may cause significant stress to users.
Challenges we ran into
The project held key concerns from the start, especially in privacy and scalability. Thus we were careful to structure our product where all parties would consent to the service, to work in tandem to better understand and communicate with each other. We knew that if our users could feel safe with their privacy protected, then they would be comfortable using our service to have a two way smart communication with their friends and family.
Accomplishments we're proud of
We are proud to showcase a strong product that will support an unvaluable human trait, communication. With an intuitive UI, safe web scraping, understandable text-to-speech, and a strong user experience, our platform can help guide our users into a safer, closer, and open social interactions.
What we learned
Aside from the technical skills in web scraping, natural language processing, and image recognition, we learned to work as a team. This project was essential to work together, because we wanted to build something meant for interpersonal conversations. Thus we learned the importance of collective input, listening to each other, and working towards building a great product. We are happy to have worked together to build this product.
What's next
Speakeasy may find a future in wearable and consumer electronics, where we can transcribe live conversations, input the text into our engine, and return an AI speech of audible warnings or hints during a conversation through an earpiece or even AR glasses integration.
Speakeasy
AI-powered conversation companion that helps you stay connected with friends and family through empathetic social insights.
https://github.com/user-attachments/assets/cc916d80-f4f4-4028-985b-839f5bc1187a
Overview
Speakeasy bridges the gap between our online and offline social lives by intelligently analyzing public social media data to provide contextual conversation guidance. Never miss important updates about your friends again.
Features
- Smart Social Monitoring: Analyzes public posts from platforms like X (Twitter) with mutual consent
- Empathetic Insights: Identifies important life events from casual posts to serious milestones
- Conversation Guidance: Provides real-time suggestions and empathetic warnings during interactions
- Multi-modal Analysis: Processes both text and images from social media posts
- Privacy-First: Requires explicit consent from all parties before data collection
How It Works
- Users install the desktop application and grant permission to analyze their social media
- AI processes public posts to identify key life events and emotional context
- During conversations, the system provides subtle guidance and context about recent updates
- Future integration with wearables will enable real-time conversation assistance
Tech Stack
- Frontend: HTML, CSS, JavaScript (built with Cursor)
- Backend: Python, PyTorch, Playwright
- AI/ML: Natural language processing, image recognition
- Audio: LMNT for text-to-speech synthesis
- Hardware: Custom heartbeat sensor for stress detection
Future Vision
- Wearable integration (earpieces, AR glasses)
- Real-time conversation transcription and analysis
- Stress monitoring through biometric sensors
- Enhanced privacy controls and data encryption
Analysis
View
Metric
- 6
Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.
Technology
- CSSIn code
- ExpressIn code
- HTMLIn code
- JavaScriptIn code
- PythonIn code
- ReactIn code
6 of 6 appear in the indexed code.
AI coding agents
No AI coding agent signals were found in this repository.
Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.
Codebase size
Source size
183 KB
Source files
30
Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.
Repository
spicylemonade/Speakeasy
47 files · 2.4 MB · @ e5d10a9
Structure
Interface
10 files · 21%Screens, components and styles rendered to the user.
Application logic
16 files · 34%Domain rules, services and shared utilities.
Supporting
Layers are inferred from where files sit in the tree, not from reading the code. A project that names its directories unconventionally will read oddly here — open the file browser to check anything the diagram implies.
Languages
- JavaScript67%
- CSS16%
- Markdown9%
- Python6%
- HTML1%
Share of indexed source by file size. Binary and vendored files are excluded.
Dependencies
CareAware/frontend/package.json
npm · 10- @testing-library/dom
- @testing-library/jest-dom
- @testing-library/react
- @testing-library/user-event
- axios
- react
- react-dom
- react-router-dom
- react-scripts
- web-vitals
CareAware/backend/package.json
npm · 6- @serialport/parser-readline
- axios
- cors
- express
- serialport
- +1 more
Declared in the repository’s manifests at the indexed commit. A declared package is not proof it is used, and runtime dependencies are listed first.
This project’s features have not been analysed yet.
Export this project's context (description, README, evidence, key source files) to chat with an AI agent elsewhere.