Connector

REST API → BigQuery

The most common shape. A vendor or internal REST API, a handful of endpoints, and data that needs to be in BigQuery reliably rather than occasionally.

What it costs

Two kinds of line. The connector is the technical connection to this source, charged once. The data models are the datasets you actually asked for, priced by type.

Build: Base Connector · Standard source · incl. 3 data models
Price: See the price list

Data models customers ask for on this kind of source

Three Simple ones come with the Base Connector. These are what tends to get added.

Data model	Typical type	API resources it usually needs
Employees	Standard	/employees, /employee-fields, /departments, /locations
Customers	Standard	/customers, /customer-tags
Orders	Advanced	/orders, /order-lines, /products
Invoices	Advanced	/invoices, /invoice-lines, /payments

Illustrative, not a quote: the type is assessed per customer from the API's own documentation. Note the last column — those resources are included in the data model price, not charged separately, and their raw data lands in your warehouse too.

What decides the type

How the resources in a dataset relate to each other
Whether the API can return only what changed
How much history has to be loaded
Nested structures that have to be flattened

What you get

A Cloud Run Job on a schedule, in your GCP project
Incremental sync with checkpointing, so an interrupted run resumes
Historical backfill executed in full, as a separate re-runnable job
Monitoring, alerting and a runbook naming the real failure modes
Source code, infrastructure definitions and documentation, in your accounts

What we deploy

A container on Cloud Run Jobs, triggered by Cloud Scheduler, reading credentials from Secret Manager and writing to BigQuery. No orchestrator, no platform, nothing running when it is idle.

Cloud Scheduler ──▶ Cloud Run Job ──▶ BigQuery
                         │
      Secret Manager ────┘
                         ▼
              Cloud Logging + Cloud Monitoring

If your data is large or the API is slow, the backfill runs as a separate parallel job through Cloud Storage. That is a decision we make from the API’s rate limits, not a decision you have to make.

What lands

Raw data, typed and named, in a dataset you own. We do not model it, and we recommend you do that in dbt where your analysts can see and test the logic. Transformation inside a connector is invisible from the warehouse and it is the main way a fixed-price build turns into a project.

When this is the wrong answer

If Fivetran, Airbyte or the vendor’s own integration already covers this API, use it. It will be cheaper than a custom build and somebody else maintains it. We will tell you if that is the case.

Get an estimate for rest api → bigquery

Send the documentation link. You get a complexity tier, a recommended architecture, a fixed price and a delivery window — free, usually the same day. If a vendor connector would serve you better, we will say that instead.

Get an estimate