Everything in Asemic runs against your own warehouse: the model is derived there, queries execute there, and results are served from there. Connecting it is step one because nothing else can happen without it. It is also the reason your raw event data never leaves your infrastructure.

On a new project, the Data Model page shows a connection strip (“Step 0 — connect your warehouse to load tables”) and a single card: “Connect your database”. Select “Connect warehouse…” to open the “Connect your warehouse” dialog, then pick your provider.

BigQuery

Enter the “GCP Project ID” of the project that holds your event data, then choose an auth method:

Service account key (recommended)

Paste the service account’s JSON key. This is the right choice for anything durable: the connection is a workload identity that belongs to the project, not to a person, so it keeps working when individuals change teams or lose access.

The service account needs read access to your raw event datasets, and write access to the dataset where Asemic builds the model (you choose it during event setup).

Google account (OAuth)

Select “Sign in with Google” and authorize. No key file; this is the fastest path when you’re evaluating. The signed-in account needs BigQuery access to the project’s data. For a connection your team relies on, prefer a service account.

Snowflake

Enter “Username”, “Password”, and the “JDBC URL” of your Snowflake account.

Test, then save

Select “Test connection” first (it verifies that Asemic can reach the warehouse with the credentials given), then “Save connection”. Testing before saving turns a mistyped project ID into a ten-second fix instead of a confusing failure later, during event discovery.

Once saved, the Data Model page loads your datasets and the next card appears: Set up events.