Bagofwords

Bag of words is an open source AI analyst that connects a language model of your choice to the data your company already stores, then answers questions and builds reports from it. Rather than a chat box bolted onto a warehouse, it combines structured planning, observations and explicit logic so an analysis can be traced back through the steps that produced it, refined in conversation and shared as an interactive report. The platform is organized around context. Teams write custom instructions and AI rules, put them through approval workflows, and ingest existing definitions from dbt, Tableau, LookML, code and AGENTS.md files, so business logic such as how ARR or daily active users are calculated lives in one governed place. The system proposes improvements to that context based on feedback and user requests, and an observability layer reports context coverage, agent accuracy, clarification counts, errors and user feedback alongside detailed traces of each agent decision and the tables it queried most. Connections span data warehouses and services such as Salesforce and AWS, any LLM provider with the ability to switch models, and enterprise controls including single sign on, role based access and self hosting. A Data MCP server exposes the same analytics agent to other AI clients. The vendor publishes no price list. The site offers a self serve deployment it says takes under two minutes and a demo booking form for teams that want to talk first, but there is no pricing page and no plan names or amounts appear anywhere in the site text or its structured data. Anyone evaluating cost will need to contact the company directly.