Skip to content

Agencies

Bulk runs

Run a list of up to 500 websites in one job and get them back sorted worst-first, each with a short report.

Before you start

  • Any account can use bulk runs, agency or not.
  • You need a list of website addresses: a CSV or text file, or a list to paste.
  • Your wallet must cover the whole list. The job won't start until it does.

The Bulk run page is in no menu. Agency accounts open it with Run many sites on Prospecting. Any account can find it by typing "Bulk run" in the search palette (⌘K).

Start a job

  1. Open Bulk run.
  2. Type a name in Name this job, such as "Hotel prospects, October". It is optional.
  3. Choose a file under CSV or text file, or paste your list into the box below it.
  4. Click Start scanning.

The job page opens and fills in as each site finishes.

How your list is read

  • One website per line. Extra columns are ignored, so a CRM export works.
  • The first web address in each row is used.
  • A website listed twice is run once.
  • Rows past the 500th are left out.
  • If you add a file and paste a list, the pasted list is used.

If a site on the list blocked automatic reading on an earlier run, a yellow banner names the extra charge first. Click Start scanning again to go ahead.

What each site costs

Each site gets a short run, not a full one. It asks Perplexity 3 prompts, reads the site, runs the website checks and scores the answers. It writes no fixes.

What it does Where the price comes from
3 prompts to Perplexity Included in the estimate on the card
Writing the 3 prompts Prompt generation, 10 minimum
Reading the site Website crawl and Website AI read
The website checks Website checks
Scoring the answers Model answer scoring

The pill on the card shows you an estimate per site. That estimate leaves out the Website crawl and Website checks rows, so each site costs you those two rows more.

A site that blocks automatic reading adds a Blocked site visit. A failure on our side is refunded. Card prices are on Prices and limits.

Read the results

The job page shows how many sites have finished and refreshes itself until all are done. The table is sorted worst-first.

Column What it shows
Website The site checked
Score Its Overall score, out of 100
Visibility The share of its answers that named the site
Status Ready, Failed, or still in progress
Report Open report for a ready row, or the start of the error for a failed one

Each report is a short version: 3 prompts, with the website checks and no written fixes. It carries your white label if you set one. The Your jobs card on Bulk run lists your 20 latest jobs.

Export the results

Click Download CSV on the job page. The file has one row per site, worst-first: the website, its score, its Visibility, its status and its report link.

Check it worked

Rows turn Ready as they finish, and each opens its report. On Bulk run, Your jobs shows the job with its finished count.

FAQ

Do bulk sites become brands?

No. They don't appear on Workspaces or Pipeline. To track one, add it as a brand. On an agency account, you can set up a pitch for it from Prospecting.

Does a bulk job stop my other runs?

No. Bulk rows don't count toward the 3 runs you can have queued or running at a time.

Can I score a site on more AI models?

Not in a bulk run, which always asks Perplexity 3 prompts. Add the site as a brand for a full run on the AI models you choose.

Open bulk runs