Connect Snowflake

Axiome now connects to Snowflake. Dashboards, filters, alerts and reports, AI summaries, forecasts, dbt imports and Axi all run on it, with SQL Axiome compiles for Snowflake itself — and without a password: Axiome signs in with a key pair it keeps for your workspace.

What changed

• One script to set it up. The connection form writes it with your names and Axiome's public key: a small warehouse that suspends after a minute idle, a role that only reads, and a service user carrying the key. Run it once in Snowsight, test, tick your schemas.
• No private key to handle. Axiome generates the key pair once per workspace and keeps the private half; your Snowflake user carries the public one. Regenerate replaces it, and says first which connections it will stop until their users carry the new key.
• Test never wakes your warehouse. Neither does reading your schema: both use Snowflake's metadata. Test says what is wrong, most fundamental first, with the line of SQL that fixes it.
• Your warehouse bill, in view. Every query Axiome runs is tagged with what it served — a dashboard, an alert, Axi — so you can find it in your query history. Repeats are answered from Snowflake's result cache, and on a Snowflake connection the alert editor says what a 15-minute check costs before you save it.
• Your names, read the way you read them. ORDERS and ORDER_DATE show as orders and order_date, while Axiome's SQL keeps the exact names. Keys and comments declared in Snowflake become joins and descriptions, and timestamps with a time zone are read in UTC.

📖 Connecting Snowflake · What it costs · Rotating the key · Import from dbt on Snowflake