Wren AI

Wren AI is built by Canner, Inc., a company founded in 2018 that first developed Canner Enterprise, a context layer connecting warehouses, lakes and databases behind one governed interface. Canner is a distributed team serving enterprise customers in North America, Europe, the Middle East and Asia-Pacific, and Wren AI is now its main focus.

Wren AI is an open-source GenBI data agent: users type business questions in plain English and receive SQL, charts and live dashboards, with each answer showing its reasoning and query. A semantic layer written in the MDL modelling language, drafted by an onboarding agent and approved by the data team, keeps definitions consistent. Row- and column-level access rules apply to people and to AI agents connecting over MCP. It connects to Snowflake, BigQuery, Databricks, PostgreSQL, SQL Server, Oracle, MySQL and 20+ more sources, and runs as Wren AI Cloud, in a private VPC or fully on-premise. Named users include Phison and Nextlink.

• Text-to-SQL agent: Writes and runs SQL for each question and shows its reasoning.
• GenBI Apps: Generates live dashboards from a single prompt that can be filtered and drilled into.
• MDL semantic layer: Metrics, relationships and definitions approved by the data team.
• Access controls: Row- and column-level security, field masking and audit logs for users and agents.
• Multi-channel access: Ask in Wren, Slack or Microsoft Teams, or connect Claude and ChatGPT via MCP.
• Flexible deployment: Wren AI Cloud, own VPC or fully on-premise, with an open-source core.

A Vietnamese electronics manufacturer can deploy Wren AI on-premise so plant managers ask about yield and downtime in natural language. A bank can let branch managers query loan data with row-level security. A retail chain can generate sales dashboards per region from simple prompts.

Business users get answers and dashboards from plain-language questions without waiting for analysts, while the SQL behind each answer stays visible for verification. Data stays in place with no ETL or migration. Fully on-premise deployment suits regulated sectors such as manufacturing and banking; Phison runs it on-prem. An open-source core and cloud plans from $179/month with a 7-day trial lower the cost of evaluation.

Generative BI is spreading as companies want managers to query data in natural language, but trust depends on a governed semantic layer and visible SQL. In Asia, manufacturers and banks often require on-premise deployment, which favours open-source GenBI that can run inside their own infrastructure.

Canner, Inc. was founded in 2018; its privacy policy lists an address in Lewes, Delaware, USA. The team is distributed and serves customers in North America, South America, Europe, the Middle East and Asia-Pacific. The Wren AI GitHub project has over 17K stars, and the site shows a SOC 2 badge.