Chat with PDF

Summary

The Chat with PDF Agent lets you ask questions about a PDF and get answers grounded in the document, with a page citation for every claim. It indexes the PDF into a vector store, retrieves only the relevant chunks per question, and can also generate comprehension quizzes from real passages. Reach for it to turn manuals, papers, and reports into something you can query.

Install

bunnpmpnpmyarn
$ pnpm dlx shadcn@latest add @agentcn/flue/chat-with-pdf
Copy

Composition

Copyagents/
└── chat-with-pdf.ts      # Conversational Q&A agent
workflows/
└── chat-with-pdf.ts      # Index-then-answer, typed cited output
tools/
├── index-pdf.ts          # Chunk, embed, and store a PDF
└── search-docs.ts        # Retrieve relevant chunks for a query
lib/
└── vector-store.ts       # libSQL vector store + embeddings helpers

Customization

• Swap the vector store. lib/vector-store.ts wraps libSQL — replace it with Pinecone, Qdrant, Chroma, or pgvector behind the same functions.
• Tune chunking. Adjust chunkText's size and overlap for your documents.
• Reshape the answer. Edit the Answer valibot schema in the workflow.
• Swap the embedding model. Change the model in lib/vector-store.ts.