Wren AI

Wren AI is an open-source generative business intelligence platform designed to translate plain-English questions into governed SQL queries, charts, and dashboards. It is structured as an open-core system with a public context engine and CLI available on GitHub, alongside commercial cloud and self-hosted editions. The platform aims to bridge human users and AI agents by providing shared semantic modeling, execution-level data policies, and integration capabilities across numerous data systems.

• Text-to-SQL generation from plain-English questions with execution-level governance
• Automated creation of interactive dashboards, charts, and spreadsheets with drill-down, roll-up, and global filters
• Agentic mode supporting sandboxed, multi-step reasoning, tool execution, and persistent agent memory and skills
• Open MDL semantic modeling where definitions are stored as versioned, git-native files
• Unified row-level and column-level security policies applied identically to human users, REST APIs, and MCP clients
• Support for over twenty data sources including BigQuery, PostgreSQL, ClickHouse, and Amazon Redshift
• Embedded analytics options including iFrame threads, a REST API for custom UX, and MCP connectivity for tools like Claude and ChatGPT

Pros

• Provides a unified semantic context layer so both human users and AI agents reference identical definitions and policies
• Enables rapid deployment of conversational analytics and dashboards without requiring users to write manual SQL queries
• Offers flexible deployment models ranging from multi-tenant cloud to private cloud, air-gapped on-premise infrastructure, and self-hosted enterprise setups
• Maintains transparency by keeping the core context engine open source on GitHub and utilizing version-controlled git-native files for semantics

Cons

• Free tier is restricted to 2 projects, 2 members, and 10 tables per project, with free monthly credits expiring at the end of each month
• Advanced governance features such as row- and column-level data control, advanced RBAC, and auditing logs require upgrading to the Enterprise Cloud or Enterprise Plus plans
• Self-hosted Enterprise Plus deployments require an annual contract starting at 10 concurrent sessions rather than individual user licensing
• Specific integrations beyond the listed connectors and supported external tools like dbt should be verified against current platform documentation

Wren AI fits data teams, enterprises, CEOs, and analytics leaders who want to empower both internal human users and AI agents with self-serve, governed business intelligence. It also suits developers seeking an open-source context engine and CLI to build custom conversational data agents. It is a poor fit for organizations that require a closed-box proprietary analytics platform without git-native semantic modeling or those looking for entirely free unlimited production environments without project or table caps.

Wren AI uses a freemium model with both cloud and self-hosted plans. The Free tier costs zero dollars per month, includes twenty monthly credits, two projects, two members, and ten tables per project. Essential Cloud costs one hundred seventy-nine dollars per month billed annually, includes thirteen thousand two hundred annual credits, unlimited dashboards, and the embedded AI API. Enterprise Cloud costs five hundred fifty-nine dollars per month billed annually, includes twenty-four thousand annual credits, row- and column-level data controls, MCP integration, and advanced auditing logs. Enterprise Plus is a self-hosted option requiring an annual contract starting at ten concurrent sessions, with pricing available upon contacting sales, and is licensed by simultaneous active sessions rather than named users. An open-source version is also available for developers at no cost.

Unlike proprietary AI analytics platforms that operate as hidden black boxes, Wren AI utilizes an open-core model where the underlying context engine is public on GitHub and semantic definitions live directly in your own git repository. This allows human analysts and AI agents to share the exact same version-controlled semantic models, file-system instructions, and execution policies. Furthermore, it supports dual operating modes for both interactive human data discovery and autonomous agent-driven multi-step reasoning via MCP and REST APIs.

Wren AI works with over twenty data sources, specifically naming BigQuery, PostgreSQL, ClickHouse, and Amazon Redshift. It integrates with external tools like dbt and connects with AI clients via MCP for tools such as Claude and ChatGPT. It also provides embedded deployment capabilities through iFrame threads and REST API endpoints.

Wren AI is worth trying for data teams and developers seeking an open-core generative business intelligence platform that serves both humans and AI agents. The generous free tier allows users to test agentic and classic modes, but projects are capped at ten tables. Teams requiring production-level security, row-level data control, and higher credit limits will need to step up to Essential Cloud or Enterprise Cloud. Organizations needing air-gapped or on-premise self-hosted deployments should evaluate the Enterprise Plus tier.