> ## Documentation Index
> Fetch the complete documentation index at: https://docs.findtempo.co/llms.txt
> Use this file to discover all available pages before exploring further.

# BigQuery

> Connect your BigQuery warehouse so Tempo can read call and CRM data your team already replicates there.

If your team replicates call and CRM data into BigQuery, Tempo can read it from there. The connection is read-only. Tempo never writes to your warehouse.

BigQuery connects once for your whole organization. An organization admin connects it from the [Connections page](https://findtempo.co/app/settings/connections).

## What Tempo reads

Tempo reads the tables you select. Today Tempo uses the warehouse connection to read CRM notes replicated into BigQuery. Those notes become **CRM note** artifacts on the matching deal, grouped one per day, and feed observations like any other artifact.

> [!NOTE]
>
> Loading CRM notes from your warehouse is run by Tempo for your organization after you connect and map your note fields. Ongoing automatic sync of selected tables is not yet available. If you want warehouse notes loaded, [get in touch](https://docs.findtempo.co/support.md).

## Two ways to connect

| Method | What happens | When to use it |
| - | - | - |
| **Grant Tempo access** (recommended) | Tempo creates a reader identity for your organization. You grant it **BigQuery Data Viewer** on the datasets Tempo should read. No credential leaves your project. | Most organizations, including those whose Google Cloud policy blocks service-account key creation. |
| **Paste a service-account key** | You paste an existing service-account key file. Tempo stores it encrypted and mints access tokens from it. | When your Google Cloud organization permits key creation and you prefer to manage the account yourself. |

## Connect by granting access

**1. Open the connect dialog**

Go to **Settings → Connections** and click **Connect** on the BigQuery card. Choose **Grant Tempo access**.

**2. Copy your Tempo reader address**

The dialog shows the service account address under **Your Tempo reader**. Copy it.

**3. Grant access in Google Cloud**

In your Google Cloud project, grant that address **BigQuery Data Viewer** on each dataset Tempo should read.

**4. Verify access**

Back in Tempo, enter the **Project ID** that holds those datasets and click **Verify access**. Tempo runs a real query against one table to confirm the grant works, not just a dataset listing.

**5. Pick tables**

Tempo lists the datasets it can see. Open one and select the tables it should read, then click **Save selection**.

> [!NOTE]
>
> Verification runs a query job. If the reader can list datasets but verification fails, the reader likely lacks permission to run jobs in the project. Grant it the **BigQuery Job User** role on the project and verify again.

## Connect with a service-account key

**1. Open the connect dialog**

Go to **Settings → Connections** and click **Connect** on the BigQuery card. Choose **Paste a service-account key**.

**2. Paste the key**

Paste the whole JSON key file into the key box. Tempo fills in **Project ID**, **Service account email**, and **Private key** from it. You can also enter them by hand.

**3. Validate and continue**

Click **Validate & continue**. Tempo mints a token and lists datasets before it stores anything, so a dead key is never saved.

**4. Pick tables**

Select the tables Tempo should read and click **Save selection**.

## Permissions needed

| Role | Where | Why |
| - | - | - |
| **BigQuery Data Viewer** | Each dataset Tempo reads | Read table data and metadata |
| **BigQuery Job User** | The project | Run the read queries Tempo issues |

Tempo needs no write role. With the grant method, no key is created for the reader identity; Tempo authenticates by impersonation.

## Change the selection

Open the BigQuery card and click the table picker again. Saving a new selection adds tables you picked, keeps the ones you kept, and disables the ones you removed. A deselected table is never deleted from Tempo's configuration, so re-selecting it later does not re-read from the beginning.

## Disconnect

Go to **Settings → Connections**, find BigQuery, and click **Disconnect**. With the grant method, Tempo disables its reader identity. With the key method, Tempo deletes the stored key. Remove the grant on your datasets to revoke access completely.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.