Bag of words: The Open-Source AI Analyst

Bag of words is an open-source AI analyst designed to connect any LLM to any data source with centralized context management, trust, observability, and control. Its core capability is built around a Knowledge Harness, which provides the context that makes the agentic loop useful: scoped connections, tables and semantic objects, files, tools, instructions, identity, permissions, and prior evidence. Agents can work across databases, services, APIs, tables, semantic objects, documents, and files. The product supports ingesting from dbt, Tableau, code, AGENTS.md, and more, with AI continuously maintaining and monitoring your context. AI will auto-suggest improvements and updates to your context based on feedback and user requests. Teams can create custom instructions and AI rules with approval workflows. Evals and feedback close the self-improvement loop, so teams can review and promote better instructions rather than giving an agent unchecked authority to change its own rules. Bag of words is used for business analysis (revenue and financial analysis, operating reviews, self-serve exploration) and technical investigation (production root-cause analysis across observability systems, services, and databases). It provides high-level performance metrics including context coverage and AI accuracy. The product targets business analysts, data teams, and engineers, has a medium learning curve, and is released as open source.