Get TikTok video on-screen text
Extracts the text shown ON a TikTok post itself: the overlay captions creators put on screen, which the regular caption and transcript never carry. Two layers are merged: native text-sticker data (present when the creator used TikTok's built-in text tool) and AI OCR of the post's images, which also catches text burned in with editors like CapCut that no data provider returns. What the OCR layer reads depends on the post. On a video, the cover frame only, so text that appears later in the video is not read. On a photo post (pass its /photo/ link), every slide in order, up to the first 10; a longer photo post is read to slide 10 and the response carries a warning giving its full slide count. texts is the deduplicated union in reading order; native_texts and ocr_texts expose each layer separately and source says which layers produced data. On a photo post, images lists each slide read with its own texts and its 0-based index, so you know which slide said what. A post with no on-screen text returns an empty texts array (that is a real answer, not an error). Costs 5 credits for a video or a photo post of any length; auto-refunds when the post lookup or the whole extraction fails.
Request
- Base URL: https://www.socialcrawl.dev/v1
- URL: https://www.socialcrawl.dev/v1/tiktok/video/screen-text
- Auth: API key in header x-api-key
Query parameters
Full URL of the TikTok video or photo post (a /video/ or /photo/ link)
Headers
Send no-cache to bypass the response cache and force a live fetch. Billed at the normal endpoint cost; the fresh result is written back to cache for the next caller. Only the no-cache directive triggers this. See the Response Schema guide for details.
Optional UUID that makes the request safely retriable. The key is reserved atomically per account before billing. A concurrent retry receives 409 IDEMPOTENCY_IN_PROGRESS with a Retry-After of 1 second while the original is young, growing to a quarter of the time it has been running (at most 30); a crashed reservation expires after 5 minutes. A replay keeps the cached payload immutable except for billing metadata: credits_used becomes 0, idempotent_replay becomes true, and credits_remaining is refreshed to the current balance. A known current balance appears in both the body and X-Credits-Remaining header; no balance row resolves to 0. On a transient lookup failure, body credits_remaining is null and X-Credits-Remaining is omitted. Completed outcomes expire after 24 hours. If the claim store is unavailable, the request fails closed before billing. GET endpoints do not combine idempotent replay with Server-Sent Events: use Accept: application/json when sending this header. The unsupported key-plus-SSE combination is rejected before billing.
Response
Successful response
Example response
{
"success": true,
"platform": "tiktok",
"endpoint": "/v1/tiktok/video/screen-text",
"credits_used": 1,
"credits_remaining": 4999,
"request_id": "req-abc123",
"idempotent_replay": true,
"meta": {
"hint": {
"code": "search_multi",
"message": "You searched \"matcha\" on 3 platforms one call at a time; /v1/search/multi returns each platform's own search rows in one call at the same price per platform, and a platform with no rows is not charged.",
"method": "GET",
"path": "/v1/search/multi?query=matcha&platforms=instagram,tiktok,youtube"
}
}
}