Skip to content
Lecture Revision Lab

Step 1 · Knowledge base

Choose what the lab reads

Start with three chapters of an open statistics textbook, or add your own notes. Files are parsed, chunked and embedded in your browser, and the embeddings are cached so nothing is embedded twice.

Demo textbook

Statistics, three chapters

Barbara Illowsky and Susan Dean · OpenStax · CC BY 4.0. Converted to plain text with exercises removed; embeddings precomputed with sentence-transformers.

In use

Your files

Upload your own notes

PDF, TXT or Markdown. Parsed and stored in this browser only (IndexedDB), never uploaded to a server.

Upload only material you have the rights to use, and no personal data. If you add an API key, the passages you retrieve are sent to that provider when you generate.

Drop files here, or

Up to 15 MB each. Scanned PDFs need OCR first.

Chunking preview

How text becomes passages

The 2024 prototype stored each file as a single row, and the model only reads the first 256 word pieces of each. Chunking packs whole sentences into passages with a little overlap. These settings apply to your uploads; the demo embeddings were precomputed at 900 characters with 180 characters of overlap.

Mode
900
180