Text extract urls
Exacttext_extract_urlsDeterministic and authoritative — the canonical answer for this input.
Pull all http/https URLs out of free text, in order of appearance. Trailing sentence punctuation is trimmed. Processing multiple inputs? Use `text_extract_urls_batch` to do up to 1000 in one call.
Parameters
valuestringrequiredThe free text to extract values from.
Call it
curl -X POST https://api.snipget.ai/v1/text/extract/urls \
-H "Authorization: Bearer sk_live_..." \
-H "Content-Type: application/json" \
-d '{"value":"Docs at https://acme.com/start and https://acme.com/api."}'The Batch tab sends up to 1000 inputs in one call via /v1/text/extract/urls/batch.
Example response
{
"input": "Docs at https://acme.com/start and https://acme.com/api.",
"is_valid": true,
"values": [
"https://acme.com/start",
"https://acme.com/api"
],
"count": 2
}FAQ
What does this do?
It pulls every http and https web address out of a block of text, in order, and counts them. Handy for pulling links from a message, document, or pasted paragraph.
Does it find links without http, like www.acme.com?
No. It only matches full http:// and https:// URLs. Bare domains without the scheme are not captured.
Does it clean up trailing punctuation?
Yes. A period or comma at the end of a sentence is trimmed, so 'https://acme.com/api.' comes back as 'https://acme.com/api'.
Does it check that the links work?
No. It extracts the addresses as written; it does not visit them or confirm they load. That would need a separate request.
Is the result stable?
Yes. It is deterministic and authoritative: the same text always returns the same URLs in the same order.