Ingest a Document
Choose a sample document to load into the pipeline. This is your raw source — the text that will be chunked, embedded, and made searchable.
Pipeline Progress
0/5 complete0%
Load document
⟳ running
Select a source and load the raw text content into the pipeline.
Split into chunks
pending
Apply a chunking strategy to split the document into retrievable segments.
Generate embeddings
pending
Convert each chunk to a vector using an in-browser embedding model.
Search vector store
pending
Embed your query and rank chunks by cosine similarity.
Generate answer
pending
Assemble retrieved context into a prompt and produce a grounded answer.