MCP_AI_FILE_FINDER

• 

  UI with simple greeting

• 

  AI response to asking for a list of folders available in the google drive

• 

  Prompting the AI to show the contents inside a folder

• 

  AI ability to summarize the contents of a file

• 

  Code file that was summarized in the example (The chatbot was able to make an accurate summary of the source code)

Inspiration

We were tired of wasting time digging through folders and repositories to find files we knew existed. We wanted an AI that could understand what we meant, not just what we typed.

What it does

MCP_AI_FILE_FINDER lets you search files using natural language. It understands context, meaning, and code intent to find files across local and cloud sources instantly.

How we built it

We combined a FastAPI backend and OpenAI embeddings for semantic search.

Challenges we ran into

Balancing speed with accuracy, parsing mixed file types such as code, PDFs, and notes, and managing large embedding indexes.

Accomplishments that we're proud of

We built a search tool that feels intelligent, understands developer intent, works across formats, and scales efficiently.

What we learned

We learned to handle asynchronous file ingestion, and design an AI that balance performance and interpretability.

What's next for MCP_AI_FILE_FINDER

We plan to add voice and chat-based querying, better cloud integrations, and contextual file previews powered by multimodal models.

Built With

• fastapi
• fastmcp
• html
• html/css
• javascript
• mcp
• openai
• python