Ingestion

Upload and validate documents before processing begins.

Parser

Extract structured text, tables, and metadata from raw files.

Chunker

Split documents into overlapping chunks optimised for retrieval.

Embedder

Generate vector embeddings for each chunk using the selected model.

Vector Store

Index embeddings in a vector database for fast similarity search.

Retriever

Find the most relevant chunks for a given query using hybrid search.

Generator

Synthesise a cited answer from retrieved context using an LLM.