For AI

Answers

How do I export comments from WeChat Channels videos?

Short answer

Pass the video's numeric ID to okdata's WeChat Channels video comments: about 25 top-level comments a page for $0.015. For the replies under a comment, send the comment ID and the video ID to WeChat Channels comment replies, $0.015 a page. Four pages of comments plus the replies to five top comments cost $0.135 per video.

Comments on WeChat Channels (视频号), WeChat's short-video feed, can only be read by scrolling in the phone app. There is no export and no web page to copy from. Brands that want to know what viewers say under a video, about a launch, a complaint or a rival, pull the comments into a table and analyse that. The comments endpoint pages through the top-level comments of a video by its numeric ID, each with likes and a reply count; the comments with many replies are then expanded one by one.

The calls

  1. Get the numeric video ID. The comment endpoints take only the long numeric ID, such as 15052232991067605751. From a share link, call video detail once with url=. A search result's export/… ID goes through video detail too. Videos from an account's video list already carry the numeric ID.
  2. Page through the comments. WeChat Channels video comments returns each comment's id, text, likes, replies (the reply count) and published_at. Pass page.next_cursor back as cursor and stop when page.has_more is false.
  3. Expand the replies. For comments whose replies is above zero and worth reading, call WeChat Channels comment replies with the comment ID as id and the video ID as post_id. Each reply's parent_id is the comment it belongs to; join on it when you merge the tables.

A real run

On November 20, 2026 we read a video that 瑞幸咖啡 (Luckin Coffee)'s official account posted on November 11; video detail showed 144 comments. The first page brought 25 top-level comments with has_more: true. The most-liked one:

{
  "id": "15052256239526943412",
  "post_id": "15052232991067605751",
  "parent_id": null,
  "text": "这有啥厉害的[捂脸]",
  "published_at": "2026-11-11T03:46:12.000Z",
  "likes": 758,
  "replies": 12
}

(The comment asks, roughly, "what's so impressive about this?") Its replies came back 10 on the first page, still with has_more: true, so a second page is needed for all 12. ip_location was empty on all 25 comments and 10 replies in this run.

What it costs

Per video Calls Cost
First page of comments only 1 $0.015
Four pages, replies to five top comments 9 $0.135
20 videos, four pages each 80 $1.20

Most comments have no replies, so expanding just the busy ones is enough.

Gotchas

  • Use the right ID. Neither the export/… search ID nor the v2_…@finder account ID works for comments; only the numeric video ID does.
  • Order is roughly by popularity, not strictly. In our run the second comment was a new one with zero likes. Sort by time or likes yourself once you have them all.
  • Page sizes vary. The replies page returned 10, so rely on has_more, not on the item count, to know when you are done.
  • IP location can be empty. WeChat shows a province next to many comments, but check a page before planning a regional breakdown on it.
  • Emoji are text codes. [捂脸] (facepalm) and similar are how WeChat writes its stickers in text; handle them separately in sentiment work.
  • Commenter names are personal data. Sentiment and topic analysis need the text, likes and time; leave names and avatars out unless you need them.

Sources

Where people asked this (read 2026-11-20)

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