Glasp achieves 5X cost savings in knowledge access for millions of users with Pinecone

What this post added

This post details Glasp's implementation of Pinecone for semantic search and AI clone functionality. It describes the process of converting user highlights into OpenAI embeddings and storing them in Pinecone namespaces. The post also elaborates on Glasp's chunking strategy, which is based on user highlighting behavior (2-3 sentences), and their data pre-processing steps to clean text and improve retrieval performance. The benefits of Pinecone's managed service, low latency, and ease of integration are emphasized.