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

# Connect Looker to Profound

[Looker](https://cloud.google.com/looker) is Google Cloud's enterprise business intelligence platform. Teams connect it to a data warehouse (BigQuery, Snowflake, Redshift, and others), define metrics in **LookML**, and use Looker's UI to explore data, build reports, and share dashboards.

This is **not** [Looker Studio](https://lookerstudio.google.com/) (formerly Google Data Studio). It is also **not** the same as Profound's Tableau connector, which exports Profound analytics into Tableau. Profound's Looker integration works in the opposite direction: **Profound Agents query data from your Looker instance**.

## What Looker terms mean

| Term          | What it is                                                                                                                                                              |
| ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **LookML**    | Looker's modeling language. Analytics teams use it to define dimensions, measures, joins, and business logic on top of raw warehouse tables.                            |
| **Model**     | A LookML project exposed for querying, such as `marketing` or `core`. Agent inline queries reference a model by name.                                                   |
| **Explore**   | A starting point for ad hoc analysis within a model — a curated set of related tables and fields users can query. Inline queries need both a model and an explore name. |
| **Look**      | A saved report or query in Looker. If your team already built the metric in Looker, agents can run it directly with **Run Look**.                                       |
| **Dashboard** | A collection of Looks and visualizations. Profound does not embed dashboards; agents run individual Looks or inline queries and receive the underlying data.            |

## What this integration does

Profound connects to **your** Looker instance using API credentials (`client_id`, `client_secret`, and instance URL). Once connected, Profound Agents can:

* **Run a saved Look** — execute an existing Look and return the result as JSON for downstream agent steps
* **Run an inline query** — query a specific model and Explore with custom fields, filters, sorts, and advanced options

<Note>
  The integration is **read-only** and **org-level**. Profound runs queries against Looker and returns the results; it does not create, update, or delete Looker content. One Looker connection is shared across your organization and can be selected in any agent that needs it.
</Note>

## Where you can use Looker in Profound

| Surface             | What it does                                                                              |
| ------------------- | ----------------------------------------------------------------------------------------- |
| **Profound Agents** | Add **Run Look** or **Run Inline Query** steps to pull Looker data into an agent workflow |
| **Sheets**          | Import rows from a saved Look into a spreadsheet for analysis or batch agent runs         |

Both surfaces use the same org-level Looker connection and the same read-only query capabilities.

## Prerequisites

Before connecting Looker to Profound, ensure:

* You have access to a Google Cloud Looker instance (for example, `https://your-company.cloud.looker.com`)
* You can create or use a Looker user with API credentials
* That user has permission to access the saved Looks, models, Explores, and fields your agents will query

<Accordion title="How to create Looker API credentials">
  1. In Looker, go to **Admin → Users** and select the user Profound should use for API access
  2. In the user's **API credentials** section, create a new API key pair
  3. Copy the generated **Client ID** and **Client Secret** before leaving the page — Looker may not show the secret again
  4. Confirm the user can access the Looks, models, Explores, dimensions, and measures your agents will need
</Accordion>

## Connect Your Account

1. In Profound, go to **Account → Integrations → Looker**.
2. Click **Connect account**.
3. Enter your connection details:
   * **Looker Base URL** — your instance URL, for example `https://your-company.cloud.looker.com`
   * **Client ID** — the Looker API `client_id`
   * **Client Secret** — the Looker API `client_secret`
4. Click **Save**.

Profound validates the credentials by authenticating with Looker (Looker API 4.0) and confirming the API user. Once connected, Looker appears as an available integration in Agents and Sheets.

## Troubleshooting

<AccordionGroup>
  <Accordion title="The connection fails to save">
    Confirm the base URL is the root Looker instance URL, such as `https://your-company.cloud.looker.com`. Profound also accepts URLs that include `/api` or `/api/4.0` and normalizes them to the root instance URL.
  </Accordion>

  <Accordion title="No Looks appear in agent nodes">
    Make sure the Looker API user has access to the folders and saved Looks you expect to use. If permissions were recently changed, retry after a short delay.
  </Accordion>

  <Accordion title="A query returns a Looker permission error">
    Check the Looker user's role, model set, and field-level permissions. Profound can only run queries that the connected Looker user is allowed to run.
  </Accordion>
</AccordionGroup>