Summary

Bag of Words provides an open-source AI data layer designed to deploy an AI Analyst in under two minutes. It empowers users to connect any Large Language Model (LLM) to diverse data sources, offering centralized context management, observability, and control. The platform facilitates natural language interaction for data analysis, dashboard creation, and report scheduling through an agentic loop, supporting Text-to-SQL, Text-to-Python, and Text-to-Dashboard capabilities.

How It Works

The project functions as a flexible, context-aware analytics layer. It integrates with various databases and services, alongside any chosen LLM, enriching queries with contextual information from sources like dbt models, BI metadata, documentation, and code. An agentic loop, incorporating tool use, reflection, and reasoning, drives the AI's ability to interpret user requests, interact with data sources, and generate insights or visualizations.

Quick Start & Requirements

• Primary Install: Docker is the recommended method for quick deployment.
• Run Command: docker run -p 3000:3000 bagofwords/bagofwords (uses SQLite by default).
• Prerequisites: Docker is essential. A PostgreSQL instance can be configured for more robust deployments.
• Documentation: Links to deployment guides and advanced setup documentation are implied but not directly provided in the text.

Highlighted Details

• Enables rapid deployment of AI Analysts, promising setup in under two minutes.
• Offers broad compatibility, supporting numerous LLMs (OpenAI, Azure OpenAI, Gemini, Anthropic, vLLM, Ollama) and data sources (PostgreSQL, Snowflake, BigQuery, Tableau, NetSuite, etc.).
• Features centralized context management for instructions, dbt models, BI metadata, and code, alongside an AI-driven catalog for reusable queries and datasets.
• Provides enterprise-ready features including Role-Based Access Control (RBAC), Single Sign-On (SSO via OIDC), audit logs, and VPC deployment options.

Maintenance & Community

No specific details regarding maintainers, community channels (e.g., Discord, Slack), or a public roadmap were present in the provided README content.

Licensing & Compatibility

The specific open-source license for this project is not stated in the provided text. Compatibility for commercial use or linking with closed-source applications cannot be determined without this information.

Limitations & Caveats

Telemetry for usage statistics and Intercom chat for support are enabled by default, though both can be disabled via configuration. Some data sources may be listed as inactive or in beta. The absence of explicit licensing information is a significant caveat for adoption decisions.