POST /count_occurrences • Credits: 0.5 credits per call

Count Occurrences splits a string on a separator and counts how many times each item appears. The response holds total_items, unique_items and a results array that lists each item with its count, sorted by count. Each call costs 0.5 credits.

A workflow often holds one long string of values: tags from a scraped page, skills from a profile, one column from a CSV export. Count Occurrences turns that string into a frequency table. The endpoint splits the string on the separator that you send. It trims the whitespace around each item and drops every empty item. It then counts the items and returns them sorted by count, from the most frequent to the least frequent. Count Occurrences runs inside RichAPI and calls no external service.

CAUTION: Each Count Occurrences call costs 0.5 credits. A call that returns an empty results array also costs 0.5 credits.

Count Occurrences accepts these four parameters and no other field. If the request body holds an unknown field, the API returns status 400 with the message Unknown input field(s): <name>.

The API reference shows a default of , for separator. That default is unreachable. If the request body omits separator, the API returns status 400 with the message Missing required field: separator.

Count Occurrences returns one flat JSON object. Here is the example response for the request above, which sends the string SaaS, AI, saas, security, AI, ai, SaaS , , devops with the separator ,.

The limit parameter truncates results only. Count Occurrences computes total_items and unique_items across the whole input, before the truncation.

Count Occurrences applies these rules in order:

This table walks the example input SaaS, AI, saas, security, AI, ai, SaaS , , devops through those rules. The separator is , and case_insensitive is true.

Nine segments enter, and eight items are counted. The eighth segment holds one space, so Count Occurrences drops it. total_items is therefore 8, not 9.

Four distinct items remain, so unique_items is 4. The counts are saas 3, ai 3, security 1 and devops 1. Rule 6 breaks the two ties: ai comes before saas, and devops comes before security.

Count Occurrences is case insensitive by default. The case_insensitive parameter defaults to true, so SaaS, saas and SAAS collapse into one item. The item value in the response is then lowercase. In the example above, the three differently cased SaaS segments produce one result object with a count of 3. To count each spelling on its own, set case_insensitive to false. The item values then keep their original case.

Count Occurrences counts no overlapping match, because separator is a delimiter and not a search term. The endpoint splits the string at every separator and counts what remains between the separators. The separator itself is never counted.

Two examples make the rule concrete.

To count how often one term appears in a block of text, split the text into words. Set separator to one space, then read the count of that term from results.

Keyword scoring is the common use of Count Occurrences. You take the text of a lead, count the words, and read the count of the keywords that matter to you. Follow these steps:

This request scores one profile summary of a lead:

The response holds the five most frequent words:

The summary scores ai 2, saas 2 and security 2. A lead score of 6 follows from the sum of those three counts.

Two details change a word count. Count Occurrences splits on the separator alone, so punctuation stays attached to the word. The last word of the summary is therefore controls. with a full stop. A double space between two words creates an empty segment, and Count Occurrences drops that segment before the count.

Count Occurrences costs 0.5 credits per call. At $0.02 a credit, the rate on the 1,000 and 5,000 credit packs, one call costs $0.01. Larger packs cost less per credit. The credit count does not change with the length of the input or with the number of items in the response.

The keyword scoring flow above costs 1.5 credits per lead, or $0.03. One Web Scraper call costs 1 credit, and one Count Occurrences call costs 0.5 credits.

RichAPI charges no credit for a 400, 401, 402, 403, 404, 405, 429 or 500 response. A 400 response names the field and lists the fields that Count Occurrences accepts:

No. Count Occurrences treats separator as a delimiter, not as a search term, and it never returns a count for the separator itself. To count one term inside a block of text, set separator to one space and read the count of that term from results.

Count Occurrences takes no search term at all. One call counts every distinct item in the input at once, so one response already holds the count of each term that you care about.

No. The input parameter takes a string. If you send an array or a number, Count Occurrences returns status 400 with the message Invalid value for input: expected string.

Set limit to 10. The results array then holds the 10 items with the highest counts, and total_items and unique_items still describe the whole input.

Yes. Count Occurrences costs 0.5 credits for every 200 response, and an empty results array is a 200 response. A 4xx or 5xx error costs no credit.

Count Occurrences reads case_insensitive as a JSON boolean. A quoted string such as "false" is a non-empty value, and Count Occurrences treats it as true. Send false without the quotation marks.

Yes. The n8n operation name is count_occurrences. The MCP tool name is count_occurrences. Both take the same four parameters as the HTTP endpoint.