Loading...
An AI-powered learning workspace featuring context-aware subchat branching, real-time SSE streaming, Tavily web search, and an integrated rich-text notes system with vector semantic search.
Clear Learn is a production-ready AI chat and learning platform built to solve context fragmentation while studying complex topics. It enables users to branch main conversations into side-panel subchats with context snapshotting, capture live selections into rich-text notes, perform 1,536-dimensional vector embedding semantic searches over personal notes, and search the live web via Tavily tool calling.
Real-time Server-Sent Events (SSE) streaming pipeline
Context-aware subchat branching with parent snapshotting
Floating text selection menu to save quotes into notes
Non-blocking 1,536-dim vector embeddings & semantic search
Automatic AI Web Search tool calling with Tavily
Split-view Quill rich-text notes with Markdown export
URL state sync without reloads (?chat=, ?subchat=, ?note=)
Designed a context-aware Subchat Branching system that snapshots parent messages up to the branch point into an isolated side panel with compact response constraints.
Offloaded embedding generation (text-embedding-3-small) to non-blocking background workers using setImmediate() with text hashing to skip redundant API calls.
Built a 3-step SSE pipeline that emits custom status event signals (e.g. searching web) prior to token streaming.