Kuaishou (China)

Kuaishou Video Search API

Search Kuaishou (China) videos by keyword and get 20 ranked results a page — caption, author, play / like / comment / share / collect counts, duration, publish time and cover image — as JSON, with no promo cards or user rows mixed in. Costs 10 credits per page; Kuaishou is priced per endpoint (1, 2, 10 or 40) and all four search endpoints are the 10. Cursor-paginated, and roughly eight times faster than the mixed Search endpoint for the same price.

POST/api/v1/kuaishou/search/videos

Authorizations

Authorizationstringheaderrequired

Bearer authentication header of the form Bearer <token>, where <token> is your Scavio API key (e.g. Bearer sk_live_your_key).

Body

application/json

keywordstringrequired

Search keyword, 1 to 200 characters. Kuaishou is a Chinese-language platform: Chinese queries return the deep catalogue, and a Latin-script query returns whatever Kuaishou itself matches, which is usually much thinner. Hashtag text works as a plain keyword, but the Hashtag Feed endpoint is the dedicated, 1-credit way to read a tag.

Example: 人工智能

cursorstring

Opaque pagination cursor. Omit it for the first page, then pass back the pcursor value from the previous response. The response field is called pcursor, not next_cursor. Do not send recoPcursor — that is a separate cursor for Kuaishou's recommendation stream. A pcursor of "no_more" means there are no further pages.

Example: 1

Request

from scavio import ScavioClient

client = ScavioClient(api_key="sk_live_your_key")

# 10 credits PER PAGE - cap the loop.
cursor, rows = None, []
for _ in range(5):
    page = client.kuaishou.search_videos(keyword="人工智能", cursor=cursor)
    for item in page["data"].get("mixFeeds", []):
        if item.get("itemType") != 5:      # skip promo / non-video cards
            continue
        feed = item["feed"]
        rows.append((feed["caption"], feed["view_count"], feed["share_info"]))
    cursor = page["data"].get("pcursor")   # NOT next_cursor / recoPcursor
    if not cursor or cursor == "no_more":
        break

print(len(rows), "videos")

Response

"result": 1,

"pcursor": "1",

"recoPcursor": "no_more",

"llsid": "2006155792367070114"

"response_time": 2560,

"credits_used": 10,

"credits_remaining": 4753

}

from scavio import ScavioClient client = ScavioClient(api_key="sk_live_your_key") # 10 credits PER PAGE - cap the loop. cursor, rows = None, [] for _ in range(5): page = client.kuaishou.search_videos(keyword="人工智能", cursor=cursor) for item in page["data"].get("mixFeeds", []): if item.get("itemType") != 5: # skip promo / non-video cards continue feed = item["feed"] rows.append((feed["caption"], feed["view_count"], feed["share_info"])) cursor = page["data"].get("pcursor") # NOT next_cursor / recoPcursor if not cursor or cursor == "no_more": break print(len(rows), "videos")