Index what was read, not what was visited
The extension tracks reading time and scroll depth and only sends pages the user actually engaged with, which keeps noise out of the index.
Case study 02
Semantic search over the pages you actually read, with a two-stage retrieval pipeline.
Browser history search matches titles and URLs, but people remember content: "that ML article from last week". HindSite indexes the pages you actually read and lets you search them in natural language.
Capture
Chrome extension
reading time, scroll depth
Engagement filter
index only pages actually read
Index
FastAPI ingest
Cohere embeddings
1024-dimensional
PostgreSQL + pgvector
reads
Query
Query router
semantic search or live tab-switch
↺ open-tab path
Stage 1: similarity
pgvector shortlist
Stage 2: rerank
cross-encoder over the shortlist
Result or tab switch
voice or text
Architecture as built. Dashed lines are feedback paths.
The extension tracks reading time and scroll depth and only sends pages the user actually engaged with, which keeps noise out of the index.
A bi-encoder shortlist from pgvector over 1024-dimensional Cohere embeddings is re-scored by a cross-encoder, so the more expensive model only sees a handful of candidates.
A query router decides whether a request is a semantic search or a live tab switch, so switching to an open tab never has to touch the vector index.