For AI

Answers

How do I tell video notes from image notes on Xiaohongshu?

Short answer

Not from the link: both kinds open at the same /explore/<id> address. Read the type field instead, which okdata returns as video or image on every note, or filter Xiaohongshu note search with type=video or type=image for $0.02 a page. Note detail ($0.02) gives all images of an image note and only the cover of a video note.

Xiaohongshu has two kinds of notes. An image note (图文) is a carousel of pictures with a title and text; Xiaohongshu's own data calls it a "normal" note. A video note is one video with a title and text. In the app the difference is obvious, but in data it is easy to miss: both open at xiaohongshu.com/explore/<24-character id>, both have the same like, save and comment counts, and both come with images, because a video note has a cover. Code that downloads "all images of a note" quietly saves a single cover for every video note, and an analysis that compares image and video performance is wrong if the two were never split. The fix is to read the note type, which every note carries.

Where the type is

Where Image note Video note
Link /explore/<id> /explore/<id> (no difference)
type in okdata data image video
images every image, at the largest size served the cover only
Video file URL n/a not returned
Likes, saves, comments, shares yes yes
metrics.views null null

Three ways to split them

  1. Filter the search. Xiaohongshu note search takes type=video or type=image (default all), the same choice the app's search filter offers between 视频 and 图文. Use it when you want only one kind for a keyword.
  2. Read type on a note. Note detail returns type for a single ID or link.
  3. Read type in a creator's list. Account notes returns type for each note, so you can work out a creator's mix, say 70% image notes, before you brief them.
Task Calls Cost
One page of video notes for a keyword 1 $0.02
5 pages of each type for the same keyword 10 $0.20
A creator's last 20 notes with their types 1 $0.02

Example: compare image and video notes on one keyword

import os, statistics, requests

URL = "https://api.okdata.online/v1/xiaohongshu/notes/search"
H = {"Authorization": f"Bearer {os.environ['OKDATA_API_KEY']}"}

for kind in ["image", "video"]:
    body = requests.get(URL, headers=H, timeout=60,
                        params={"query": "防晒霜 推荐", "type": kind, "sort": "popular",
                                "fields": "id,type,metrics.likes,metrics.favorites"}).json()
    notes = body["data"]["items"]
    saves = [n["metrics"]["favorites"] or 0 for n in notes]
    print(kind, len(notes), "notes, median saves:", statistics.median(saves) if saves else None)

Limits

  • No video file URL. A video note gives its cover in images; the video itself is not returned.
  • type describes the note, not the creator. To find creators who mostly post one kind, PGY's creator search has its own type filter; see PGY image and video note prices.
  • Counts are the same fields for both kinds, so compare saves and comments; there is no play count. Downloading every image of image notes is covered in Xiaohongshu note images.

Sources

  1. Spider_XHS: download code branches on the note type 'video' (issue #3)github.com · read
  2. XHS-Downloader README: separate switches for image (图文/图集) and video worksgithub.com · read

Where people asked this (read 2026-11-12)

Updated

Connect once, then just ask.

Every new key comes with $0.10 to try it. One sentence connects your agent.

Start free
Connect your AI

Paste this into your agent

Read https://okdata.online/agents.md and set up okdata for me; my key is in the OKDATA_API_KEY environment variable.

Claude Code, Cursor, Codex and other agents that can open a link set themselves up from it.

MCP serverhttps://api.okdata.online/mcp