About the service
An analyst's workbench where data arrives as a file or a live database connection and questions are asked in plain language. Julius writes its own Python and R, runs the statistics, draws the charts and explains the result with citations back to the source rather than handing over a bare table. It connects straight into warehouses — Postgres, BigQuery, Snowflake, MySQL, SQL Server — and builds a semantic layer over the schema so it does not have to sift through thousands of tables per question. Findings export as a presentation, report or dashboard, recurring analyses run on a schedule, and you can feed in your own context through a data dictionary and custom agents trained on your tables. Under the hood are frontier models: OpenAI's GPT-5.6, Anthropic's Claude Sonnet 5 and Claude Fable 5, alongside Julius's own models.