Arsalan, founder of LiveDocs, has rebuilt his YC W22 startup from a basic analytics tool into an AI-native notebook that treats data work as a living system. The platform now features a reactive dependency graph where each cell recalculates only when its dependencies change, eliminating the maintenance headaches of traditional notebooks.

Unlike static dashboards or linear notebooks, LiveDocs allows users to mix SQL, Python, charts, and text in documents that stay synchronized automatically. The system pushes queries down to connected warehouses like Snowflake and BigQuery rather than copying data out, running locally on DuckDB and Polars for smaller datasets. Every result remains inspectable and reproducible.

The AI agent operates inside the notebook itself, not as a separate chat interface. It can plan multi-step analyses, write and debug code, spawn specialized sub-agents, and browse documentation when needed. Teams can publish analyses as interactive apps similar to Retool, with real-time collaboration allowing multiple users to edit simultaneously. LiveDocs targets messy analytical questions that don't fit neatly into dashboards, with pay-as-you-go pricing starting at $15 per month and a free tier for testing.