What is canTrust?
canTrust is a chat-first data studio. You ask in plain language; a governed AI Analyst reads your warehouse, shows its work, and returns dashboards, reports, and transformations you can re-run. No SQL and no pre-built dashboards required.
How it works
Connect a database, warehouse, or spreadsheet. The Analyst inspects your schema, plans a multi-step analysis, writes and runs SELECT-only queries, generates visualizations, and explains its reasoning — all in one conversation. Every step is visible in the Glass Box trace, so you can verify the logic before you act on the result.
Every data touch is governed by default: queries are read-only and RBAC- and PII-aware, and any write or DDL requires human approval. Dashboards and reports are specs that re-run, transformations are versioned snapshots, and every agent action is audited.
Connect your data
Snowflake, BigQuery, Databricks, Postgres, MySQL, SQL Server, Vertica, and Supabase — or upload CSV, Excel, and JSON files. Every connection is read SELECT-only.
Learn moreLet the Analyst work
A role-based agent plans the analysis, writes and runs the queries, and synthesizes an answer — streaming its reasoning so you can catch a wrong step early.
Learn moreTransform as code
Describe a change in plain language. canTrust previews the generated code on a safe clone, snapshots first, and applies it in a transaction — with one-click rollback.
Learn moreStay governed
SELECT-only access, RBAC- and PII-aware queries, ephemeral isolated execution, and data that never leaves your infrastructure for model training.
Learn more