Skip to main content
The Looker Studio connector pulls your Promptwatch data into Looker Studio (formerly Google Data Studio), so you can build dashboards that blend AI visibility with your other marketing data. The connector reads through the Promptwatch API and authenticates with a project API key, so one data source maps to exactly one project; an agency sets up one key and one data source per client.

​
Step 1: create a project API key

  1. Go to Settings → API Keys.
  2. Give the key a name, for example “Looker Studio - Client X”.
  3. Keep the key type on Project Key (the option labeled “Access to a single project only and Data Studio integration”) and select the project you want to report on.
  4. Keep the permission on Read-only. The connector only reads data, so there’s no reason to hand it a full-access key.
  5. Click Create Project API Key. The new key is copied to your clipboard automatically; in the keys table it shows masked, and you can reveal or copy it again later.
The Create New API Key form in Promptwatch settings with the Project Key type selected, showing access to a single project and Data Studio integration with read-only permission. API access is a plan feature: if your plan doesn’t include it, the create button is disabled with a note to upgrade, and existing keys stop working after a downgrade (they start working again if you upgrade back). Use a Project Key, not a Master Key. Master keys span all projects and require an X-Project-Id header on every call; the project key carries its project by itself.

​
Step 2: add the connector in Looker Studio

  1. Open the Promptwatch connector in Looker Studio.
  2. Authorize the connector with your Google account.
  3. Enter your project API key and connect.
  4. Create the data source, then add it to a new or existing report.
To report on several projects in one report, repeat both steps per project and blend the data sources in Looker Studio.

​
What data you can pull

The connector reads the v1 API catalog, per project. It covers the core metrics; app-only surfaces like shopping, ads, and content are not exposed:
  • Visibility: your brand visibility score as a time series, grouped by day, week, or month, filterable by model, monitor, or a single prompt.
  • Sentiment: sentiment scores over time, plus the sentiment distribution across responses.
  • Share of voice: your share of brand mentions as a time series, plus a comparison of the current period against the previous one.
  • Responses and mentions: total responses, brand mentions over time, and your top competitors by mentions.
  • Citations: citation analytics, self-citation frequency, cited domains over time, citation rank analysis, your top cited pages, and which sources and domains each model cites.
  • Competitor heatmap: the model-by-brand visibility matrix.
  • Prompts and monitors: the prompt and monitor lists behind the numbers, useful as report dimensions.
  • Crawler and visitor analytics: AI crawler activity trends and top crawled pages, plus visit totals and trends from visitor analytics.

​
How fresh is the data

The API serves reads from a short server-side cache; the window varies per endpoint, from a couple of minutes for visibility and sentiment up to longer windows for slow-moving data like monitors and models. Looker Studio adds its own data freshness interval on top, which you control per data source. The result is current enough for reporting; it’s not a live-refresh feed.

​
Troubleshooting

  • “Invalid API key”: the key was mistyped or has been deleted. Copy it again from Settings → API Keys, or create a new one; deleting a key breaks every data source built on it.
  • “API access is not available on your current plan”: your organization’s plan no longer includes API access. Upgrading restores existing keys without recreating them.
  • The numbers are for the wrong website: the key is project-scoped, so the data source shows whichever project the key was created for. Create a key under the right project and reconnect.
  • Requests failing under heavy use: API calls share your organization’s rate limits across all keys, enforced per second, minute, and hour. A report with many charts refreshing at once can hit the short windows; the response says when to retry, and the rate limits page lists the numbers per plan.

​
Other ways to get data out

Use the connector for dashboards that stay up to date on their own. For a formatted weekly deliverable, scheduled PDF reports email themselves; for one-off raw data, use CSV exports; for ad-hoc questions in your own AI tools, use the MCP server. The integrations overview lists everything side by side. After you connect, sanity-check one number you know, like this week’s visibility score against the dashboard, before you send the report to anyone.