欢迎关注我们的最新动态。
更多内容请见底部「阅读原文」。
' DIGEST="Q1 has been a wild ride — here's what shipped." SOURCE_URL="https://example.com/q1-recap" # optional; populates 「阅读原文」, leave empty to omit PAYLOAD=$(jq -nc \ --arg title "$TITLE" \ --arg author "$AUTHOR" \ --arg thumb "$THUMB_MEDIA_ID" \ --arg content "$CONTENT_HTML" \ --arg digest "$DIGEST" \ --arg source_url "$SOURCE_URL" \ '{articles: [{ article_type: "news", title: $title, author: $author, thumb_media_id: $thumb, content: $content, digest: $digest, content_source_url: $source_url, need_open_comment: 0, only_fans_can_comment: 0 }]}') curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/draft/add?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$PAYLOAD" | jq # → {"media_id": "DRAFT_MEDIA_ID"} ``` Tell the user: *"Draft created. Open https://mp.weixin.qq.com → 草稿箱 to review and tap «发表» from there."* — that's the safest path because the user controls the publish click in WeChat's own UI. ### List drafts (newest first) ```sh curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/draft/batchget?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw '{"offset": 0, "count": 20, "no_content": 1}' \ | jq '.item[] | {media_id, update_time, title: .content.news_item[0].title}' ``` ### Get the full content of one draft ```sh DRAFT_MEDIA_ID="..." curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/draft/get?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg m "$DRAFT_MEDIA_ID" '{media_id: $m}')" | jq ``` ### Update an existing draft ```sh curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/draft/update?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg m "$DRAFT_MEDIA_ID" --arg t "Updated title" --arg c "new body
" --arg th "$THUMB_MEDIA_ID" ' {media_id: $m, index: 0, articles: { title: $t, content: $c, thumb_media_id: $th }}')" | jq ``` ### Publish a draft to «发表记录» — DESTRUCTIVE, confirm before calling ```sh # Eats one of the monthly publish slots. Always echo back to the user # what's about to go live and require an explicit "yes" confirmation # in conversation BEFORE invoking this. curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/freepublish/submit?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg m "$DRAFT_MEDIA_ID" '{media_id: $m}')" | jq # → {"errcode": 0, "msg_data_id": ..., "publish_id": "..."} ``` Then poll publish status (publishing is async, takes ~5-30 seconds): ```sh PUBLISH_ID="..." curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/freepublish/get?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg p "$PUBLISH_ID" '{publish_id: $p}')" | jq '{publish_status, fail_idx, article_id, article_url: .article_detail.item[0].article_url}' # publish_status: 0 = success, 1 = publishing, 2 = original-check-failed, # 3 = failed, 4 = published-but-removed, 5 = unverified-removed ``` When `publish_status == 0`, return `article_detail.item[0].article_url` to the user — that's the canonical https://mp.weixin.qq.com/s/... URL. ### List already-published articles ```sh curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/freepublish/batchget?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw '{"offset": 0, "count": 20, "no_content": 1}' \ | jq '.item[] | {article_id, update_time, title: .content.news_item[0].title, url: .content.news_item[0].url}' ``` ### Send a customer-service message (one specific follower, 48h window) ```sh OPENID="oXXXXXXXXXXXXXXXXX" # the recipient follower's openid TEXT="Hi! Your subscription has been renewed. Thanks for sticking with us." curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/message/custom/send?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg u "$OPENID" --arg t "$TEXT" ' {touser: $u, msgtype: "text", text: {content: $t}}')" | jq ``` `errcode 45015` = recipient hasn't messaged the account in the last 48h — explain that to the user; there's nothing the API can do about it. To send an article card via customer message, change to `msgtype: "mpnews"` with `mpnews: {media_id: "..."}` (use `material/add_news` to first create a permanent news media_id). ### Pull follower-growth + read stats ```sh BEGIN="2026-04-25" END="2026-05-01" # max 7-day window for getusersummary # New / unsubscribed / cumulative followers per day curl -sS -X POST \ "https://api.weixin.qq.com/datacube/getusersummary?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg b "$BEGIN" --arg e "$END" '{begin_date: $b, end_date: $e}')" \ | jq '.list[] | {date: .ref_date, new: .new_user, lost: .cancel_user, source: .user_source}' # Article reads per day (max 3-day window for getuserread) curl -sS -X POST \ "https://api.weixin.qq.com/datacube/getuserread?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$(jq -nc --arg b "$BEGIN" --arg e "$BEGIN" '{begin_date: $b, end_date: $e}')" | jq ``` `datacube/*` requires an authenticated (已认证) account with the data-cube permission. ### Manage the bottom menu ```sh # Read current menu curl -sS "https://api.weixin.qq.com/cgi-bin/menu/get?access_token=${WECHAT_ACCESS_TOKEN}" | jq # Replace the menu (max 3 top-level buttons; each top-level button can have ≤ 5 sub-buttons) MENU=$(jq -nc '{ button: [ {type: "click", name: "今日推荐", key: "DAILY_REC"}, {name: "更多", sub_button: [ {type: "view", name: "官网", url: "https://example.com"}, {type: "view", name: "联系我们", url: "https://example.com/contact"} ]} ] }') curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/menu/create?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw "$MENU" | jq # → {"errcode": 0, "errmsg": "ok"} ``` ### Get the follower list (paginated by next_openid) ```sh NEXT="" curl -sS "https://api.weixin.qq.com/cgi-bin/user/get?access_token=${WECHAT_ACCESS_TOKEN}&next_openid=${NEXT}" \ | jq '{total, count, openids: .data.openid, next: .next_openid}' # Loop: pass the returned `next_openid` as NEXT until count < 10000. ``` For each openid, batch-fetch profile (≤ 100 per call): ```sh curl -sS -X POST \ "https://api.weixin.qq.com/cgi-bin/user/info/batchget?access_token=${WECHAT_ACCESS_TOKEN}" \ -H "Content-Type: application/json; charset=utf-8" \ --data-raw '{"user_list": [{"openid": "OPENID1", "lang": "zh_CN"}]}' \ | jq '.user_info_list[] | {openid, nickname, subscribe_time, tagid_list}' ``` ## HTML rules for article `content` The `content` field in `draft/add` accepts a *subset* of HTML, not full web HTML: - Allowed: ``, ``, ``, ``, ``, ` `) — no ``, `
`, ``–`
`, `
/
/
`, `