A product analyst needs to segment customers by purchase behavior across three years of transaction data. She's working in a Jupyter notebook, changes a filter in cell 3, then manually reruns cells 4 through 12 to see updated results. Takes forever. Livedocs handles this with reactive notebooks that automatically update downstream cells when you change anything upstream. Ask a question in plain English, get SQL queries built, charts rendered, and analysis complete in seconds.
The system covers twenty different analysis types. Sales trend analysis. Customer segmentation. Revenue forecasting. Churn prediction. A/B test analysis. Cohort analysis. Anomaly detection. Time series work. Marketing ROI breakdowns. Customer lifetime value calculations. Website traffic analysis. All through an AI agent that translates questions into executable analysis.
A marketing manager wants to understand which campaigns drive the best ROI. She connects Google Drive and her company's Postgres database, asks "which channels generated the most revenue per dollar spent last quarter," and gets back visualizations with drill-down capability. The data cleaning pipeline handles messy inputs automatically. The interactive dashboard builder lets her share findings with stakeholders who can explore the data themselves without writing code.
A supply chain analyst runs inventory optimization models every Monday morning. With scheduling on the Pro plan, he sets notebooks to refresh automatically with weekend sales data. Terminal access lets him run custom scripts when standard analysis templates don't quite fit. Real-time collaboration means his team lead can jump into the same notebook, add comments, and suggest adjustments without version control headaches.
The free plan works for solo exploration. You get three notebooks, basic datasources like CSV and Excel, and 4 GB RAM with half a virtual CPU. There's a $5 AI credits cap with no option to add more. That's enough for learning or light personal projects. Builder costs fifteen dollars monthly, bumps you to five notebooks and 8 GB RAM with 2 vCPUs, includes five dollars in AI credits then switches to pay-as-you-go. Still solo only. Pro runs thirty dollars monthly, adds twenty per extra team member, removes the notebook limit, doubles computing power to 16 GB RAM and 4 vCPUs, enables all datasources including Snowflake, Databricks, BigQuery, ClickHouse, and Motherduck. Scheduling, terminal, and collaboration features live only on Pro. Annual billing knocks twenty percent off.
A financial analyst working with proprietary on-premise databases can't use cloud-hosted tools. Enterprise pricing gets custom deployment, SSO, custom machine profiles, and dedicated support. A freelance data consultant juggling multiple client projects hits the three-notebook ceiling on free almost immediately.
The reactive notebook approach matters most when iteration speed determines whether insights arrive useful or obsolete. Traditional notebook users who've built elaborate run-order systems might find the automatic update behavior disorienting at first. The AI agent shines with straightforward analytical questions but won't replace deep statistical expertise for complex modeling work.
Teams already locked into specific BI platforms with embedded workflows won't want to migrate. Solo practitioners on tight budgets face a choice between severe notebook limits on free or stepping up to paid plans. The builder tier doesn't add team features, which feels like an odd middle ground. No mobile app means analysis happens at your desk, not during commutes or from job sites.
