你下周最好的帖，可能藏在这周收到的某条评论里。评论变内容 Prompt 把粉丝信号变成稳定的选题源，避免空白脑暴的冷启动风险。下面 15 个 Prompt 覆盖全流程：提取选题、DM 追问、把评论作为故事火花、为每个平台出版本。
适合哪些场景
评论活跃的创作者、每周几百条评论的品牌社媒负责人、跑多账号的 KOC 运营、把 DM 变成编辑 pipeline 的社群运营。
什么时候不建议这样写 Prompt
互动很低的账号（每周不到 50 条评论）——信号池太小。粉丝群体不能被引用（私密社群、敏感话题）也不要用。
Prompt 结构公式
评论变内容 Prompt 一定要带这六个要素：
- 角色：让 AI 扮演谁（小红书 KOC / TikTok 脚本写手 / 个人品牌策略师 / 社群运营）。
- 上下文：平台、领域、受众画像、账号体量、语气——任何会改变效果的信息。
- 目标：一个具体交付物——一个 hook、一句 caption、一支 60 秒脚本、10 条回复变体、一段 bio。
- 限制：字数、禁用词、本地化语感、算法信号、tag 数量、语气规则。
- 输出格式：编号选项、A/B 变体、可直接粘贴的块、JSON、或带标签的段落。
- 示例 / 信号：1-2 条你欣赏的参考帖、或反例（“不要那种泛泛的 creator 腔”）。
这套 Prompt 适合用在哪
- 每周评论挖矿出选题
- 创始人 BD 把 DM 变成思想领导力内容
- 品牌社群从 UGC 出内容
- 评论→newsletter pipeline
- KOL 用粉丝引用做复盘内容
15 个可直接复制的 Prompt 模板
1. 评论挖矿选题（10 条）
每周默认——喂 50 条评论，拿 10 个选题排序。
You are a content strategist mining audience signal. Below are 50 comments from my recent posts about {topic / niche}. Cluster into 10 topic ideas ranked by signal strength. For each: (a) topic in 1 line, (b) 2-3 exemplar comments that proved demand, (c) suggested format (Reel / thread / carousel), (d) why audience would care now.
{paste 50 comments}可替换变量： topic / niche, 粘贴 50 条评论
优化建议： 若输出太浅，追加：“Prioritize comments that contain a real question, a specific pain, or a strong disagreement. Skip generic praise.”
2. 单条评论 → 帖子
Below is one comment that struck me: "{paste}". Frame it as the spark for a 300-word post. Lead with the comment (anonymized or with permission), name the broader pattern it points to, give my take, and close with a question that invites more comments.3. 反对评论 → 立场帖
A commenter disagreed with my last post: "{paste}". Write a 200-word post that holds my position respectfully, takes their critique seriously (steel-man it first), and lands on what we both agree on. Voice: confident, not defensive.4. DM 追问拿上下文
A commenter said something interesting but vague: "{paste}". Write a 40-word DM asking for one specific follow-up question that would let me turn this into a post. Polite, no pressure, no extraction-y vibe.5. 高频问题 → FAQ 帖
Below are 30 comments. Identify the 5 most-repeated questions and turn them into a single FAQ-style carousel post. Each FAQ: question (verbatim from a comment), 60-word answer, 1 visual cue.
{paste 30 comments}6. 评论 → 长贴扩写
One commenter wrote: "{paste}". Expand their point into a 7-tweet X thread under my voice. Tweet 1: lead with the commenter's line (credited). Tweets 2-6: my expansion. Tweet 7: invite further responses.7. 客户引用三件套（已授权）
A customer left a glowing comment: "{paste}". Write 3 content asset variants: (a) testimonial graphic copy ≤30 words, (b) case-study LinkedIn post (250 words), (c) 30-second Reel script in their voice. Include a permission-request DM template.8. 评论聚类 → 系列
Below are 40 comments clustered by my software into 4 themes ({themes}). Design a 4-post series — one per theme. For each post: hook, topic, format, target platform. The series should feel like one ongoing conversation with the audience.
{paste 40 comments}9. 评论引发的反主流观点
Most commenters on my recent post about {topic} agreed with me. Pick the 1 thoughtful dissent: "{paste}". Write a follow-up post that publicly engages with their dissent, names what I am updating in my view, and invites the original commenter to respond.10. Live Q&A 题库
I am running a 30-minute IG / Xiaohongshu Live Q&A. From the 50 comments below, pull the 12 best questions to answer live. Order by audience appeal (broad first, then specific). Time-box each at 2-3 minutes.
{paste 50 comments}11. 反”误解”内容
Below are 5 comments that misinterpreted my last post. Write a follow-up clarifying post that is generous (assumes readers wanted to engage), not defensive. Lead with what I could have said more clearly, then the actual point.
{paste 5 comments}12. 评论 → newsletter 板块
For my weekly newsletter, write a recurring section titled "{section name}" that pulls 3 reader comments from the past week into a curated short essay. Each comment gets 1 line of context + 50-word reflection. Voice: curator, not influencer.13. 评论触发的产品想法
Below are 20 comments hinting at things the audience wishes existed: "{paste}". Extract the top 3 product / service ideas implied. For each: pattern of demand, 2-3 supporting comments, smallest viable test I could run.14. 评论变内容审计
Over the last 4 weeks I generated content from {N comments}. Below is the list of comment-spawned posts and their performance. Audit: which comment patterns yielded the strongest posts? Which yielded duds? What signal should I weight more?
{paste posts + metrics}15. 授权 + 致谢模板
每条公开引用都用一次。
I want to quote {commenter} publicly. Write 3 permission DM variants: (a) full attribution, (b) anonymized with their preferred handle, (c) paraphrased without quote. Include a 1-line credit template I can use in the eventual post.容易踩的坑
- 没问就引用别人评论——多数地区合法，但会消耗信任。
- 只挖正面评论——反对和提问通常含更强信号。
- 同一个人话被你引 5 次——观众注意得到，会感到被消费。
- 没有 DM 追问——评论往往只是更深 insight 的表层。
- 把一条情绪评论当趋势——等到集群再做。
- 不致谢——被致谢的人会放大你，不致谢的人下次冷处理。
- 没追踪”评论→内容”的产出率——不追踪就不知道哪类信号最值钱。
优化技巧
- 每周做一次 30 分钟评论挖矿——回报是稳定的选题 pipeline。
- 每天标几条”可发”评论，免得被新评论淹没。
- 一条评论被用作内容就 DM 原作者——礼数会复利。
- 先聚类再写——一条是轶事，三条是模式。
- 每月用模板 14 审计哪类评论产出最强。
- 维护小型授权 DM + 引用致谢模板库提速。
- 把评论挖矿和日历规划绑在一起——最强信号填最高优先级时段。
FAQ
- 引用评论一定要授权吗？: 公开评论多数地区可直接引用，但信任层面：用对方名字、原话或图片之前请先 DM 征得同意。
- 几条评论才算一个 pattern？: 3 条独立评论说同一件事。一条是轶事，两条是巧合，三条才值得做一帖。
- 匿名评论也要致谢吗？: 要——可以用”某位读者上周问”这种模糊但诚实的说法，依然建立信任。
- 最好的评论来自竞争对手或黑粉怎么办？: 公开针对内容本身回应，不针对语气。处理得当时，反对者带来的内容回报往往高于赞美。
- 评论要多新？: 趋势话题 14 天以内；常青选题 60 天以内。太老的引用会显得迟到。