Synthia by NuvelaAI

Basic landing page where user can enter information about their paper
Project tab to show the paper and its specific details
Showcase of grabbed fragments of other papers that are also relevant to the user's work
A citation generator for these fragments
A semantic match accuracy detector to show how relevant these fragments are to the user's work
Latex preview of final generated paper content
PDF preview of final generated paper content

Inspiration

MCP (Model Context Protocol) seems like the next step in AI progression. https://www.anthropic.com/news/model-context-protocol. We wanted to implement this into a very practical application, that being generating and matching research paper citations and sources to inspire more accurate and relevant research papers.

What it does

Synthia provides an intuitive dashboard for configuring and managing research projects that integrate with MCP. This allows the user to describe their paper, and then semantically match it to certain fragments of candidate sources, which will be matched with and included in a final generated paper.

How we built it

Using a reactjs frontend ensures the most leisured UX/UI experience, added to an MCP implementation in the backend using python, with Cohere for embedding and Pinecone for database management.

Challenges we ran into

Ensuring the frontend design matches the general feel of what we want to accomplish, meshing the frontend and backend together, along with getting the data to actually match with summaries and paper descriptions.

Accomplishments that we're proud of

Successfully implementing a functional integration between MCP and the React dashboard, developing a scalable architecture that allows for future expansion and additional research capabilities, improving the efficiency of model management for academic research, reducing overhead for researchers, creating a clean and modern UI that simplifies project tracking and configuration.

What we learned

Best practices for structuring a full-stack application with Python and React, the importance of designing APIs that efficiently handle research data and model configurations, strategies for optimizing frontend-backend communication to minimize latency, real-world challenges in building research-focused platforms and how to address them.

What's next for Synthia by NuvelaAI

Expanding this MCP architecture to other territories, where NuvelaAI can deploy even more services and ensure the upmost efficiency and assistance to as many people as possible.