nao

nao is an open source analytics agent builder built around context engineering. A data team assembles everything the agent should know as a file system: warehouse schemas with column descriptions, previews and profiling, dbt and Looker repositories, business definitions, example queries, metric semantics, rules and MCP tools. The finished agent is then deployed as a chat interface where anyone in the company can ask questions about the data in plain English. It is aimed at analytics and data engineering teams that want a reliable self service agent rather than a generic text to SQL bot. A command line workflow covers the lifecycle: nao init creates the context project, nao sync pulls context automatically from databases such as BigQuery, from code repositories and from external sources such as Notion, nao test runs unit tests that pair questions with expected SQL and reports pass rate, answer rate, time and tokens, and nao chat launches the interface. The chat supports data stories and monitoring of user conversations, and questions can also come in through Slack, Teams, WhatsApp, Telegram, Claude, Codex and Cursor via the nao MCP server. Teams bring their own LLM key and pay only for token consumption. The Self Hosted plan is free and includes self hosted deployment, your own LLM key, full access to the context builder and chat UI, and Slack integration. Cloud, hosted on nao Cloud with preconfigured email and Slack, Teams or WhatsApp bots plus priority support, is custom priced and currently offered as early access. Enterprise is also custom and adds SSO with Okta or Microsoft, data rights impersonation from the warehouse, row level security, LLM budgets per user, a company branded UI and SOC 2 Type II reports. Because the self hosted edition keeps data and keys inside the customer's own infrastructure, it suits companies with strict data residency requirements.