> ## Documentation Index
> Fetch the complete documentation index at: https://promptwatch.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Prompt Explorer
> The Prompt Explorer generates suggested prompts from Google Search Console, the keywords a page ranks for, a seed topic, People also ask, or one of your pages, so you can grow your tracked prompt set from real search data instead of guessing.
### What is the Prompt Explorer?
The Prompt Explorer is a workbench for finding new prompts to track, built on real search data instead of brainstorming. You pick a source, Promptwatch generates a batch of suggested prompts from it, and each suggestion lands in a review table with search volume and difficulty attached. From there you decide: track it in a monitor, or dismiss it.
This matters because your metrics are only as good as the prompts behind them. A monitor full of prompts nobody asks produces clean charts about questions that don't exist. The Explorer anchors every suggestion to something observable: a query you rank for, a question Google surfaces, or a page you published. See [choosing the right prompts to track](/docs/academy/right-prompts-to-track) for the strategy side.
Click **Explorer** in the left rail. It is a standalone page titled **Prompt Explorer**, with no inner navigation. You can also reach it from the **Create Prompts** menu on the Prompts page, or through cmd+K (Ctrl+K on Windows). Viewers can see the page, but the controls are disabled: generating and importing need an editor role.
### The five sources
Each source is a tab at the top of the page:
* **Google Search Console**: generates prompts from queries your site already ranks for, so you extend proven search demand into AI answers. It reads the last 30 days and up to 100 queries. It needs a [Search Console connection](/docs/academy/google-search-console) and a property linked to the project: if the account is not connected, the tab shows **Connect Google Search Console** (or **Reconnect Google Search Console** when the connection errored); if the account is connected but no property is chosen, a **Set up Search Console for this project** button takes you to the project's Search Console settings.
* **Ranking keywords**: you enter a page URL or a domain, like "example.com/product" or "example.com", and Promptwatch uses the keywords it ranks for in Google (up to 100) as the seed. No Search Console connection needed. If Google has no rankings for that page yet, the tab says so.
* **Manual Search**: you enter a seed topic, like "best crm software", and Promptwatch expands it into natural-language prompts around that topic.
* **People also ask**: you enter a Google query and Promptwatch turns Google's "People also ask" questions for it into prompts. These are questions real users ask, phrased the way they ask them.
* **From website page**: you paste one of your URLs, Promptwatch crawls its title, description and content, then generates the prompts that page could be the answer to. Useful right after publishing something new. If the page can't be read, you get "We couldn't read that page. Check the URL and try again."
Every tab lets you set the target country (it defaults to your project's country), which drives search volume and the People also ask results. All tabs except People also ask also let you set the **Prompt type** and **Intent type** for the generated suggestions, defaulting to Organic and Commercial; on People also ask the generator assigns them. See [working with prompts](/docs/academy/working-with-prompts) for what the types and intents mean. Google Search Console, Ranking keywords and Manual Search write suggestions in your project's language; People also ask and From website page write them in the primary language of the country you picked.
Click **Find Prompts** to generate. Each run adds up to 10 suggestions. On an empty table you'll see a "Discovering prompts" loader, then a toast with how many prompts were discovered (or "Prompts are up to date" when nothing new came out). If you have never generated anything, the empty state reads "No discovered prompts yet" with a **Discover prompts** button that builds a first batch from sources you already have connected.
### Reading the suggestions table
Suggestions from all sources collect in one table below the tabs, newest first, so the batch you just generated sits on top; new rows stay highlighted until you act on them. Each row shows:
* **Prompt**: the generated prompt text.
* **Type**: compact letter badges for the type and intent the suggestion will carry into tracking; hover for the full names.
* **Volume**: the search volume of the suggestion's primary keyword, so you can judge how much demand sits behind it.
* **Difficulty**: how competitive that keyword is in search.
* **Date**: when the suggestion was generated.
Prompts you already track are filtered out automatically (matched on the prompt text), so the table never suggests duplicates of your existing setup. Once a project has 50 active suggestions, generating more fails with "You have reached the suggested prompt limit. Track or dismiss suggested prompts before generating more."
### How to turn a suggestion into a tracked prompt
1. Click **Track** on a single row, or tick several rows with the checkboxes (the header checkbox selects all) and click **Add to Monitor**. Both open the same dialog.
2. In the **Import Prompts to Monitor** dialog, pick the **Monitor** that should own the prompts. Only active monitors are listed, and the first one is preselected. Under **Prompts to Import (N)**, remove any you don't want with the trash icon.
3. Click **Add to Monitor**. A toast confirms how many prompts were added.
If the project has no active monitor, the dialog says "No monitors available" and offers a **Create Monitor** button; see [creating a monitor](/docs/academy/creating-a-monitor). Closing the dialog after removing prompts asks "Discard import?" first. If the import would exceed your plan's active prompt limit, the upgrade dialog opens instead.
Not sure which monitor? See [what is a monitor](/docs/academy/what-is-a-monitor) and [organizing your setup](/docs/academy/organizing-your-setup): the short version is that a prompt should join the monitor matching its market and language.
Suggestions you don't want, dismiss with the trash icon on the row, or tick several and click **Dismiss**. Dismissed suggestions stay dismissed, even when a later run produces the same prompt again, so the table stays focused on what you haven't decided on yet.
### After you import
The new prompts are created inside the monitor you picked, take over that monitor's language, keep their suggested type and intent, and are queued to run immediately rather than waiting for the next daily cycle. First responses typically show up shortly after, and the prompts count against your plan's active prompt limit like any other prompt. From there they behave exactly like prompts you [added by hand](/docs/academy/adding-prompts).
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.