Arxchive

• 

  Chatting with paper

• 

  PDF viewing

• 

  Fetching papers

Inspiration

Chat with pdf apps and personal research challenges

What it does

Fetches research papers from arxiv, stages paper, chunks and adds them to snowflake table, then performs rag on the questions by fetching relevant chunks.

How we built it

we started with the snowflake rag tutorial then went from there.

Challenges we ran into

pdf embedding, rag filtering, and latency in pipeline.

Accomplishments that we're proud of

completing the app, building rag pipeline

What we learned

working in snowflake worksheet, sql, managing data pipelines

What's next for Arxchive

Adding more chat features, more interactive pdf viewing, chatting with research papers not from arxiv, and enhancing rag pipeline, pdf uploads are stored in s3 bucket

Built With

• arxiv
• mistral
• python
• snowflake
• streamlit