Investor research assistant

Research the letters.
See the evidence.

Explore Berkshire Hathaway shareholder letters from 1977 through 2025. The chatbot retrieves the most relevant passages from Qdrant, then asks an AI model to answer only from that evidence.

Private retrieval Grounded answers Visible sources

Investor research

Ask the shareholder letters

Try a question

RAG Assistant

Ask me about Berkshire Hathaway’s shareholder letters from 1977 through 2025. I’ll retrieve relevant passages first, answer from that evidence, and show you exactly which years and pages were used.

Why RAG

A chatbot that looks before it answers.

Retrieval makes a general AI model useful on private or specialized knowledge without retraining it.

Your data stays authoritative

The response is constructed from your indexed material instead of relying on broad model memory.

Evidence stays inspectable

Every response returns the retrieved chunks and their similarity scores for quick review.

Knowledge updates independently

Add or replace vectors in Qdrant without retraining the answer model or rebuilding the interface.