Jepto does not invent a fixed field list. You register one table or view, discover its columns, then map types and aggregations. Queries run live against BigQuery — no freeform SQL.
| Entity | What it contains |
|---|---|
| Table or view | One BigQuery table or view per data source. Use a view in BigQuery if you need a multi-table shape. |
| Fields | Columns discovered from INFORMATION_SCHEMA, then mapped to Jepto types (number, string, date, boolean). |
| Source config | GCP project, BigQuery dataset and table bound per client (Client Config on template sources). |
| Partition column | Date/time partition used to prune scans. Page date ranges map to this column so queries stay cost-safe. |
| Template definition | Optional shared schema for master reports, with a different project/table per linked client. |
Reports, forecasts and alerts — built on the same live connection.
Put CRM orders, pipeline or warehouse KPIs on the same client report as Google Ads and Meta.
Explore reportingTurn on Make this data source a template, then bind each client’s project and table in Client Config.
Explore template data sourcesSet monthly targets on warehouse metrics you already trust — not just platform-reported conversions.
Explore KPI forecastingAlert when a warehouse metric moves outside its normal band, the same way Jepto watches ads and analytics.
Explore anomaly detectionAsk about connected tables in plain language — read-only, no write-back to BigQuery.
Read-only analytics
Use MCP to explore the fields you registered and query live results. Jepto maps the question to declared columns and aggregations — it never sends freeform SQL from the chat.
schema()
Explore the fields, types and aggregations registered on the connected BigQuery source.
data()
Query live results from the bound project, dataset and table using declared fields only.
actions()
Not available — this is a read-only source. No write-back actions are supported.
execute()
Not available — this is a read-only source. Jepto cannot modify data in the platform.
Sign in with a Google Cloud account that can query the project, then pick one table or view.
Click Connect and authenticate with a Google account that has BigQuery access to the client’s GCP project.
Select the GCP project, BigQuery dataset and table (or view). Jepto discovers columns and suggests types.
Set aggregations, hide unused columns, optionally turn on Make this data source a template, then assign it to the client.
Jepto queries BigQuery when a report, Data Chat or MCP request runs. History is whatever the table contains; partitioned tables require a date range.
Everything you need to know about BigQuery in Jepto.
No. Jepto queries BigQuery live when you run a report, Data Chat or MCP request. It does not copy the table into a Jepto warehouse.
No. You register columns and aggregations (SUM, AVG, MIN, MAX, COUNT). If you need joins inside BigQuery, create a view there and connect that view.
Turn on Make this data source a template to share the schema across clients. Each client still binds their own project, dataset and table in Client Config.
Partitioned tables require a date filter mapped from the report range. Jepto dry-runs queries and rejects missing or oversized ranges so you do not scan the full table by accident.
No. Data Warehouse exports Jepto’s marketing data to BigQuery. This source registers a table you already own so it can be reported on next to ads and analytics.
Related Articles
Start asking your data smarter questions today. Connect your first data source in minutes.
