BambooAI

Enables users to analyze data and derive insights using natural language queries without extensive coding knowledge.

Automatically generates and executes Python code for data analysis and visualization based on user prompts.

Automatically corrects errors in generated code through an LLM-based debugging process, ensuring successful execution.

Utilizes an optional planning agent to break down complex data analysis requests into manageable, step-by-step plans.

Supports analysis of both local datasets and external data fetched from APIs and internet searches.

Allows simultaneous work with multiple datasets, intelligently analyzing, loading, and joining them for comprehensive analysis.

Integrates custom ontologies to provide the AI agents with domain-specific knowledge, improving data interpretation and code accuracy.

Stores and retrieves successful analysis solutions in a vector database, enabling the system to learn and reference past insights for future tasks.

Leverages specialized AI agents (e.g., Planner, Code Generator, Error Corrector) to handle different aspects of the analysis workflow.

Integrates with various API-based LLMs (OpenAI, Gemini, Anthropic, etc.) and local models (Ollama, VLLM) for agent configurations.

Users can provide and utilize custom YAML-based prompt templates to tailor the AI's behavior and responses.

Offers detailed parameters for controlling conversational context, search tools, planning, and web UI deployment.

Provides a Flask-based web application for an interactive data analysis experience with a visual workflow map.

Fully supports interaction and analysis within Jupyter notebooks for a programmatic user experience.

Ensures enhanced security by executing generated code in a sandboxed environment, especially when deployed via Docker.

Facilitates easy deployment through Docker for a consistent environment or via a standard Python pip package installation.