LabRuns entirely in your browser. Nothing is sent to any server.
A whole RAG in your tab
Paste a document or upload your .txt files and watch the RAG pipeline from inside, stage by stage: text gets chunked, every chunk becomes a vector, your question hunts its neighbours and the answer is assembled from the most relevant sentences, each with its citation. The part that decides what an assistant knows is exactly this.
The full version, with a generator, on HF SpacesMore experiments: