BambooAI
An LLM-driven data analyst that runs in a persistent Python kernel, writing and executing its own code one cell at a time and re-running its cited results in a fresh kernel before presenting a report.
BambooAI is an LLM-driven data analyst that works like a person stepping through a notebook: given a dataset and a question, it writes and runs Python code one cell at a time in a persistent kernel, reads the output, decides the next step, and ends with a report. Before that report is shown, the cells it cites are assembled into a script and re-executed in a fresh kernel against the original data, and the report states how many of its cited numbers reproduced.
It offers three modes: a quick lookup or definition, a deep analysis within a turn budget, and an adaptive long run with periodic self-review by a stronger model. Follow-up questions continue on the same warm kernel state, while a new question starts fresh, and a Gemini key lets it search the web for published figures to compare against.
It runs on a user's own machine with the user's own model keys, and the model-written code executes inside a container rather than under the user's account. It grew out of sports-science data analysis and applies to any tabular dataset.