Get unstuck. Fast.
Browse the docs, scan the FAQ, or talk to our team. Enterprise customers reach a named technical contact directly.
Frequently asked questions
The questions data leaders, engineers, and security teams ask most.
How do I connect a new datasource?
From the Colrows console, open Connectors and pick the engine (Snowflake, Databricks, BigQuery, Redshift, Postgres, MySQL, and 10+ more). Provide credentials and a scope. Colrows runs an introspection pass and proposes initial semantic mappings, which you can edit before publishing.
Where are my queries logged for audit?
Every prompt, compile, plan, and execution is logged with a unique query ID, the user/persona that issued it, the policies applied, and the final dialect-specific SQL that ran. Audit logs are point-in-time reproducible - you can re-run a historical query and prove it used the exact definitions in force at that moment.
Can I run Colrows in my own VPC?
Yes. Colrows deploys on-premise or into your private VPC on AWS, Azure, or GCP, licensed per tier with no vCPU or per-node charge. Common in financial services and healthcare. Reach out via contact for a deployment topology review.
How does compile-time governance differ from BI semantic layers?
BI semantic layers (Looker, Power BI) resolve meaning at presentation time, inside a single vendor. Colrows resolves meaning at compile time, across any warehouse, and enforces policy before any query touches your data. The result: governance is structural, not advisory - unauthorized queries fail compilation, not in production.
How is Colrows usage metered?
Colrows meters one thing: natural-language requests. Each question to the AI Analyst, each document question, each Semantic Resolution API request and each agent call counts once, however many internal steps it takes. Dashboards, reports and re-runs of an existing answer are not metered. Administrators are alerted at 80% of capacity, and above 110% new questions pause until the next cycle or an approved query pack - never a surprise bill.
How do I migrate from a dbt semantic layer?
Colrows can ingest dbt metric definitions and convert them into the semantic graph. Once imported, the graph evolves autonomously - drift detection, conflict resolution, and schema changes are handled by AI agents instead of by hand. Your dbt models continue to compile; Colrows just becomes the runtime above them.
Does Colrows support custom SQL dialects?
Out of the box: Snowflake, Databricks, BigQuery, Redshift, Postgres, MySQL, ClickHouse, Trino, and more. The compile pipeline emits dialect-specialised SQL per engine. For exotic in-house engines, the dialect adapter is pluggable - talk to us about contributing one.
How do I report a security issue?
Email dev@colrows.com with details. We acknowledge security reports within 24 hours and have a disclosure process for responsible reporters.
Talk to a real human.
Our team replies same-day for most questions. Enterprise customers reach a named contact directly.