Summary

Auto-Analyst is an open-source, modular AI system automating data science workflows, from cleaning and analysis to machine learning and visualization. It targets data scientists seeking an efficient, LLM-agnostic platform with a user-centric interface and the flexibility to use their own API keys, reducing vendor lock-in and costs.

How It Works

The platform employs a modular agent architecture powered by DSPy, where specialized agents handle tasks like data preprocessing, statistical analysis, machine learning, and visualization. A central planner orchestrates these agents, automatically routing queries, generating execution plans, and consolidating results. Its LLM-agnostic design allows seamless integration with various LLM APIs, and users can leverage their own API keys for cost control.

Quick Start & Requirements

A live demo is available at https://www.autoanalyst.ai/chat. Installation details for local deployment are not explicitly provided in the README, suggesting initial use via the web application. Developers can integrate custom agents by implementing dspy.Signature classes. No specific hardware, OS, or Python version prerequisites are listed for the web app.

Highlighted Details

• LLM Agnostic & BYO API Key: Compatible with any LLM API (OpenAI, Anthropic, etc.) and allows users to utilize their own API keys, avoiding vendor lock-in.
• Modular Agent System: Extensible architecture where agents (e.g., @preprocessing_agent, @sk_learn_agent) are implemented as dspy.Signature classes, facilitating custom agent development.
• Built-in Connectors: Supports data ingestion from Google Ads, Meta, LinkedIn Ads, HubSpot, Salesforce, and various SQL databases (Postgres, MySQL, Oracle, DuckDB).
• User-Centric UI: Features a chat interface and a code editor with AI-assisted edits and auto-fix capabilities.
• Planner Mode: Automates query routing, agent selection, and inter-agent workflow coordination.

Maintenance & Community

Contributions are welcomed, including adding new agents, improving UX, and submitting bug reports. Community engagement is fostered through links to Substack and LinkedIn. For enterprise inquiries or collaboration, contact is available via https://www.autoanalyst.ai/contact. The roadmap outlines goals for enhanced analysis modes, multi-file handling, and UI-driven agent creation.

Licensing & Compatibility

Auto-Analyst is released under the MIT License, a highly permissive open-source license that allows for free use, modification, and distribution, including for commercial purposes.

Limitations & Caveats

Current support for multi-CSV or multi-sheet Excel analysis is listed as a future enhancement, suggesting potential limitations in handling complex file structures. The availability of additional data connectors is noted as "upon request," implying the current list is not exhaustive. Some advanced features, like the Analytics Dashboard, are designated for enterprise use.