Use Grok Bot to Scan Following List and Filter 72-Hour Viral Posts by Views-to-Followers Ratio

Input: X following list + domain keyword → Process: Grok Bot filters by ratio/originality/type → Output: Viral post results table + content pattern summary

Who

X content creators and operators who want to learn effective posting strategies from their following list

You get

Get a viral-post list sorted by views-to-followers ratio, distinguish save-worthy vs. comment-driven content, and extract reusable content patterns

You need

Grok Bot

Steps

1. Prepare filter criteria and send the instruction

   Open a Grok Bot conversation and paste the instruction below exactly. The instruction presets a 72-hour window, originality filter, follower cap (150), ratio calculation, per-author cap (3), domain split, save/comment type tags, and more. Replace [your domain, e.g., AI/startup] with your actual domain before sending.

   Execution prompt · Source text。Grok Bot conversation input box

   Copy prompt

   帮我整理我关注的账号在最近 72 小时发的原创帖（不算转发和回复），找出浏览量高、点赞多的帖子。要求：
   1. 关注的人太多的话，先按粉丝数取前 150 个，并告诉我有多少没扫到。
   2. 除了原始浏览量，再算「浏览量 ÷ 作者粉丝数」，按这个倍数单独排一张表。
   3. 每个作者最多保留 3 条，避免大号霸榜。
   4. 单独列出跟【你的领域，比如 AI/创业】相关的帖子。
   5. 每条给出：作者、粉丝数、发布时间、浏览/赞/转发/回复/收藏、一句话摘要、原帖链接。
   6. 标出收藏多于赞的「收藏型」，和回复远多于转发的「评论型」。
   7. 总结它们为什么爆、我能借鉴哪几条。成人、猎奇、诱导互动的帖子单独标出，不要当学习对象。
   8. 只读，不要替我点赞、关注、评论或转发。拿不到的数据直接写「没拿到」，不要估。

2. Review the AI-returned filtered results and ratio table

   Grok Bot will automatically: read following list → take top 150 by follower count → filter original posts from last 72 hours → calculate views-to-followers ratio → sort by ratio → keep max 3 per author → split domain sub-table → tag save-worthy vs. comment-driven → exclude adult/gossip/engagement-bait. Wait for the results table and verify it includes: author, follower count, post time, views/likes/reposts/replies/bookmarks, summary, link, ratio, and type tag. If a filter fails (as happened with the min-likes filter in the source), have the AI report it openly and switch to a fallback (e.g., pulling per account individually); do not allow estimation or fabricated numbers.

3. Analyze viral patterns and plan content topics

   Based on the returned table, focus on "save-worthy" posts (bookmarks > likes, usually lists/playbooks/pitfall records, longer-lasting) and high-ratio posts (content that broke beyond the author's existing follower circle). Combine with the AI-summarized "why it went viral" patterns (hook opening, 72-hour hot topics, actionable takeaways, numbered lists, thread splitting, ending question) to plan your own content. The source author's specific actions: write save-worthy pitfall lists, post same-day handling notes when industry events happen, split one hot topic into 2-3 posts spaced 30-60 minutes apart, write "N things AI did for me this week" with before/after time, give clear judgments on hot AI products with 2-3 reasons.

4. Safety tip

   Safety tip: before the bot publishes publicly, review the content and confirm.

5. Note known limitations and boundaries

   Known limitations from the source: accounts with fewer than ~3,145 followers may be skipped (57 of 207 not scanned); protected accounts unreadable; max 25 posts per account (frequent posters only show latest 25); long posts truncated to ~280 characters; all numbers are snapshot at fetch time not real-time; AI may discover mid-process that some filters failed—have it report openly and adjust, never estimate. Final topic choice, opinion, and publish decision remain human; AI only searches, filters, and organizes.

帮我整理我关注的账号在最近 72 小时发的原创帖（不算转发和回复），找出浏览量高、点赞多的帖子。要求：
1. 关注的人太多的话，先按粉丝数取前 150 个，并告诉我有多少没扫到。
2. 除了原始浏览量，再算「浏览量 ÷ 作者粉丝数」，按这个倍数单独排一张表。
3. 每个作者最多保留 3 条，避免大号霸榜。
4. 单独列出跟【你的领域，比如 AI/创业】相关的帖子。
5. 每条给出：作者、粉丝数、发布时间、浏览/赞/转发/回复/收藏、一句话摘要、原帖链接。
6. 标出收藏多于赞的「收藏型」，和回复远多于转发的「评论型」。
7. 总结它们为什么爆、我能借鉴哪几条。成人、猎奇、诱导互动的帖子单独标出，不要当学习对象。
8. 只读，不要替我点赞、关注、评论或转发。拿不到的数据直接写「没拿到」，不要估。