How to Use the Looker Studio Connector
The LLM Pulse Looker Studio connector lets you visualize your AI visibility data inside Looker Studio (formerly Google Data Studio). Build live dashboards for mentions, citations, visibility, sentiment, and competitor benchmarks, then share them with your team or clients without giving them an LLM Pulse login.
With the Looker Studio connector, you can:
- Build live AI visibility dashboards in Looker Studio, refreshed from your LLM Pulse data
- Choose between four data sets: Brand trends, Prompts, Share of Voice, and Cited sources
- Filter and break down any chart by AI model, country, language, tag, brand focus, or prompt type, right in the report
- Blend AI visibility metrics with your other marketing data sources
- Share client or stakeholder dashboards without giving them an LLM Pulse login
- Start from a ready-made template instead of building from scratch
Before you start
You need two things:
- A Scale tier or higher plan. The connector authenticates with an API key.
-
An LLM Pulse API key. Open the profile menu in the bottom left of the app, go to API Keys, and create one if you haven't already. Keys start with
llmpulse_. Keep it private: anyone with the key can read your project data.
Quick start
You have three ways to get going. Most people start with the ready-made template.
- Ready-made template (fastest): Clone our pre-built dashboard, then point it at your data. Open the template
- Connector (build from scratch): Add the data source and design your own report. Add the connector
- Video walkthrough: Prefer to follow along? Watch the setup video
You can also reach all three links from Integrations in the LLM Pulse sidebar.
Set up the connector
The setup asks for two things only: a project and a data set. Everything else (dates, granularity, metrics, and filters) is chosen in the report, like any other Looker Studio data source.
1. Add the data source
Open the connector link and authorize it with your Google account if Looker Studio asks.
2. Enter your API key
Looker Studio asks for your llmpulse_ API key in its secure credentials step. Paste the key there and continue. The connector validates it before showing the settings. The key is stored as a private user credential, not as a report configuration field.
3. Pick a project and a data set
Choose the project you want to visualize, then the Data set, which decides what each row represents:
- Brand trends: metrics for your brand and every tracked competitor over time. This is what the ready-made template uses.
- Prompts: one row per prompt, with its text, link, classification, visibility, citation and sentiment metrics.
- Share of Voice: each brand's share of all brand mentions over time.
- Cited sources: the domains cited in AI answers, ranked by how many responses cite them.
To combine several data sets in one report, add the connector once per data set.
4. Connect
Click Connect in the top right. Your data source is ready, and you can drop its fields into any chart or table.
Build charts and filters in the report
- Dates follow the report. The connector always reads the window of the report's date range control (the last 28 days when the report has none).
- Pick the granularity with the date field. Use Date for daily data, Week for weekly data, or Month for monthly data. Rates such as Visibility are computed for the period you chart, so a weekly chart shows each week's real rate rather than an average of days. Charts without a date field, such as scorecards, show one total for the whole date range: counts are added up, and rates such as Visibility are computed over the whole range.
- Filter or break down by segment. AI Model, Country, Language, Tag, Brand Focus, and Prompt Type are regular dimensions. Add one to a chart to split it, or add a filter control (a drop-down list, for example) to let viewers pick values. For example, a drop-down on Tag lets viewers see one tag at a time.
- Metrics are fetched on demand. The connector only asks LLM Pulse for the metrics your charts use.
Each combination of segment values in one chart is a separate request to LLM Pulse, so a chart can split the data into at most 40 combinations (8 for Cited sources). If a chart goes over, the connector says so: add a filter on one of the segments or remove one from the chart.
Use the ready-made template
If you'd rather not build from scratch:
- Open the template.
- Click Use my own data (or Make a copy) in the top right.
- When prompted, add the LLM Pulse connector and authenticate with your API key.
- Select your project and the Brand trends data set, and the charts repopulate with your data.
The template ships with the essential AI visibility charts already built, so you get a working dashboard in a few clicks.
Available fields
Brand trends, Share of Voice, and Cited sources share these segment dimensions:
| Field | Description |
|---|---|
| AI Model | AI model that produced the answers |
| Country | Country of the prompt (ISO code) |
| Language | Language of the prompt (ISO code) |
| Tag | Prompt tag. A prompt with several tags counts once in each, and untagged prompts are left out when you chart by tag |
| Brand Focus | brand, brandother, or nonbrand. Use non_brand for fair comparisons with competitors |
| Prompt Type | informational, navigational, commercial, or transactional |
Brand trends
| Field | Type | Description |
|---|---|---|
| Date, Week, Month | Dimension | Period of the data point |
| Actor Type | Dimension | "project" (your brand) or "competitor" |
| Actor Name | Dimension | Name of the brand or competitor |
| Actor Domain | Dimension | Domain of the brand or competitor |
| Actor ID | Dimension | Internal ID of the project or competitor |
| Visibility | Metric | Share of responses that mention the brand |
| Mentions | Metric | Responses that mention the brand |
| Responses | Metric | Responses collected (see Good to know) |
| AI Visibility Score | Metric | Position-weighted visibility (higher is better) |
| Average Mention Position | Metric | Average rank of the brand among the brands named in the responses that mention it (lower is better) |
| Citations | Metric | Responses that cite the brand's URL. Includes visible citations and background source references |
| Average Position | Metric | Average rank of the brand's URL among the URLs a response cites (visible citations only, lower is better) |
| Citation Rate | Metric | Share of responses that cite the brand's URL |
| Sentiment (5 levels) | Metric | Share of the brand's sentiment verdicts at each level, from very positive to very negative |
| Net Sentiment Score | Metric | Overall sentiment balance, from -100 to 100 |
Prompts
| Field | Type | Description |
|---|---|---|
| Prompt ID, Prompt | Dimension | The prompt and its internal ID |
| Prompt Link | Dimension | Opens the prompt in LLM Pulse |
| Prompt Tags | Dimension | Tag names separated by "; " (filter with "contains") |
| AI Model | Dimension | Adding it splits each prompt into one row per model |
| Country, Language, Brand Focus, Prompt Type | Dimension | The prompt's classification |
| Responses | Metric | Responses collected for the prompt in the period |
| Mentions, Visibility | Metric | Responses mentioning your brand, as a count and a share |
| Average Mention Position | Metric | Average rank of your brand among the brands named in the responses that mention it |
| Citations, Citation Rate | Metric | Responses citing your brand's URL, as a count and a share |
| Average Citation Position | Metric | Average rank of your cited URL among the URLs a response cites |
| Sentiment Responses | Metric | Responses with a sentiment verdict for your brand |
| Sentiment (5 levels), Net Sentiment Score | Metric | Sentiment of the prompt's responses, as shares of Sentiment Responses |
Share of Voice
| Field | Type | Description |
|---|---|---|
| Date, Week, Month | Dimension | Period of the data point |
| Actor Type / Name / Domain / ID | Dimension | Same actor dimensions as Brand trends |
| Share of Voice | Metric | The brand's share of all brand mentions in the period |
Cited sources
| Field | Type | Description |
|---|---|---|
| Domain | Dimension | The cited domain |
| Responses Citing Domain | Metric | Distinct AI responses citing the domain |
| Domain Citation Rate | Metric | Share of all responses citing the domain |
Good to know
A few behaviors that trip people up:
-
Percentages use a 0 to 1 scale. Looker Studio expects percentage fields between 0 and 1, so a Visibility of 45% appears as
0.45. This is correct: Looker Studio adds the % sign automatically in charts. It is the same value you would see in the REST API (45), just on a different scale. - Sentiment fields are shares, not filters. Neutral Sentiment is the share of neutral verdicts among the responses that carry a sentiment for your brand. It never changes Responses: a prompt with 32 responses, 13 of them neutral, shows 32 responses and about 41% neutral. Use Sentiment Responses in the Prompts data set for the base count.
- Responses repeat on every actor row. In Brand trends, each brand and competitor row carries the same Responses count, so a chart that adds Responses across actors multiplies it. Filter on Actor Type = project, or break the chart down by actor.
- All competitors are included. Brand trends and Share of Voice always include your brand and every competitor in the project. Filter on Actor Name to narrow the comparison.
- Gaps in charts are normal. Prompt executions follow the tracking cadence configured for your account: daily, weekly, or monthly. Charts have no execution data on days without a run, so chart by Week or Month if your cadence is weekly or monthly.
Change your configuration later
To switch projects or data sets after setup:
- In Looker Studio, go to Resource → Manage added data sources.
- Click Edit on the LLM Pulse data source.
- Click Edit connection, update the project or data set, and click Reconnect.
Changing the data set changes the available fields, so charts that use fields the new data set lacks need updating.
Data sources created with the classic setup
Data sources added before October 2026 keep working exactly as before, with the classic setup: a report type, metric checkboxes, filters, competitors, a date range, and a granularity chosen in the data source configuration. Nothing changes in your existing reports.
In the classic setup:
- The Include ... Sentiment and other metric checkboxes only choose which metrics to fetch when a chart asks for none. They never filter responses, and they have no effect on the Prompt performance report type.
- Filters, competitors, and the date range apply to the whole data source. Choose Follow report date range to use the report's date range control.
To move a report to the new setup, add the connector again (it opens with the new setup), pick the same project and the matching data set (Brand trends for Timeseries, Prompts for Prompt performance, Share of Voice for Share of Voice, and Cited sources for Top Sources), then switch your charts to the new data source. Brand trends and Share of Voice keep the classic field names, so their charts carry over.
Troubleshooting
"Unable to fetch data"
Check that your API key is active and starts with llmpulse_, that the selected project still exists, and that your Scale tier or higher plan is current. Reconnect the data source if Looker Studio asks you to enter the key again.
"This chart splits the data into more than 40 combinations"
The chart uses too many segment dimensions at once (for example AI Model, Country, and Language together). Add a filter on one of them or remove one from the chart.
No data in the charts
Widen the report's date range, chart by Week or Month to match your account's tracking cadence, and check the report's filters on AI Model, Country, Language, and Tag.
Data looks out of date
Looker Studio caches results. Use Refresh data in the report or data source. The underlying data updates when prompt executions complete according to your account's tracking cadence.
A competitor shows no data
They may have no mentions for your selected date range or filters. Confirm the competitor exists under Preferences → Competitors in the LLM Pulse sidebar.
Frequently asked questions
Can I blend this with Google Analytics or Search Console?
Yes. Use Looker Studio's data blending and join on the Date field to combine LLM Pulse data with any other source.
Can I build calculated fields?
Yes. Create custom calculated fields in Looker Studio from any of the metrics above, for example your own ratios or combined scores.
Can I add more than one LLM Pulse data source to a report?
Yes. Add the connector once per project or data set and use them together in the same report.
Does Looker Studio's own date range control work?
Yes. The connector always fetches exactly the window selected in your report's date range control. Classic data sources do the same when their date range is set to Follow report date range.
Can viewers filter by tag or AI model?
Yes. Add a drop-down list control on Tag, AI Model, or any other segment, and viewers can switch values without editing the report.
Can I use this on a free or Starter plan?
No. The connector needs an API key, which requires a Scale tier or higher plan.
