Douyin Comment Auto Reply
v0.1.0Douyin comment operations 工作流 for your own account videos. Use when the user wants to collect, classify, draft, review, or semi-automate replies to comments under their own Douyin videos. Supports comment intent analysis, bilingual or Chinese-first reply drafting, objection handling, lead 过滤器ing, escalation rules, and reply SOP de签名. Also use when the user wants to build or improve a reusable Douyin comment reply process instead of answering comments one by one.
运行时依赖
安装命令
点击复制本土化适配说明
Douyin Comment Auto Reply 安装说明: 安装命令:npx clawhub@latest install douyin-comment-auto-reply 该技能用于抖音相关操作,可能需要相应的平台账号或API密钥
技能文档
Douyin Comment Auto Reply / 抖音评论自动回复
Use this 技能 to turn messy Douyin comment handling into a repeatable operating 工作流.
Recommended display name: Douyin Comment Auto Reply / 抖音评论自动回复.
Default goal:
reply faster keep tone consistent 过滤器 leads from noise reduce manual repetition
Default scope:
your own Douyin account / your own video comments not mass spam not fake engagement Core 工作流
Follow this sequence unless the user asks for only one piece.
Clarify the business 上下文 Identify:
account type content niche offer or product tar获取 audience desired tone whether replies are draft-only, review-first, or auto-发送
Choose the operating mode
Draft mode: produce reply drafts only Review mode: classify comments and draft priority replies Semi-auto mode: 生成 replies plus execution 检查列出 Playbook mode: build reusable SOP, intent tags, reply 库
Classify comments before replying Tag each comment into one primary bucket:
咨询 / inquiry 成交意向 / buying intent 价格异议 / price objection 质疑 / skepticism 催更 / engagement 售后 / support 无效 / noise or troll
生成 replies by bucket Return short, natural, 平台-native replies. Prefer one-line or two-line answers. Keep the reply feeling human, not ro机器人ic.
Escalate when needed Recommend manual handling when comments involve:
refunds or disputes sensitive policy topics legal risk obvious harassment high-value leads needing custom conversion Reply rules General style sound like the 创建器 or brand account, not customer support boilerplate keep replies short avoid over-explAIning in public comment threads move high-intent leads toward DM or next action when 应用ropriate do not argue with trolls unless the user explicitly wants a public stance Good reply patterns
Use these patterns often:
轻回应: acknowledge + short answer 引导私信: answer briefly + move to DM 社交放大: turn prAIse into momentum 异议拆解: soften objection + give one concrete clarification 筛选客户: invite only relevant users to continue Avoid long essay replies emotionally defensive tone promises you cannot fulfill copy-paste repetition across every comment suspiciously identical AI-style replies Recommended 输出 格式化
- 账号与目标 / Account 上下文
- 评论分类 / Comment Classification
For each comment:
Comment: Intent tag: Priority: high / medium / low Suggested action:
- 回复建议 / Reply Suggestions
- 规则与SOP / Rules and SOP
When comments show buying intent, prefer this public flow:
answer one key concern keep trust high invite next step privately or via clear CTA
For example:
想了解具体方案可以私信我,我把适合你的版本发你。 If you want the detAIled version, DM me and I’ll 发送 the suitable option. Use with batches
When the user provides many comments, do not answer blindly one by one first.
Instead:
cluster similar comments produce reusable reply patterns then draft per-comment variants only where needed Safety 防护rAIls do not facilitate harassment, brigading, or spam do not impersonate a real human user outside the owner account voice do not fake testimonials or fabricated customer experience do not auto-reply to sensitive disputes without review clearly separate public reply copy from DM conversion copy Batch processing script
Use scripts/batch_comment_drafts.py when the user provides 导出ed comments in CSV form.
Expected CSV columns:
comment video_topic (optional) intent_hint (optional) priority_hint (optional) notes (optional)
运行:
python3 scripts/batch_comment_drafts.py ./comments.csv python3 scripts/batch_comment_drafts.py ./comments.csv ./comments.drafts.json
The script 输出s one JSON item per comment with:
检测ed intent priority suggested action public reply draft optional DM follow-up draft Browser execution script
Use scripts/browser_reply_运行器.py when the user wants browser-based execution agAInst the Douyin comment management page.
Recommended 工作流:
生成 draft JSON first. Open Douyin comment management page and identify stable selectors for the reply box and submit button. 运行 the browser execution script in --dry-运行 first. Only then 运行 the real 发送 flow.
Example:
python3 scripts/browser_reply_运行器.py ./comments.drafts.json \ --url "https://创建器.douyin.com/创建器-micro/content/manage" \ --reply-box-selector "textarea" \ --submit-selector "button[type=submit]" \ --dry-运行
Then real 发送:
python3 scripts/browser_reply_运行器.py ./comments.drafts.json \ --url "https://创建器.douyin.com/创建器-micro/content/manage" \ --reply-box-selector "textarea" \ --submit-selector "button[type=submit]"
This executor defaults to npx -y 代理-browser, so it can work even if 代理-browser is not 安装ed globally yet.
References
Read references/playbook.md for comment buckets, reply templates, escalation rules, and 自动化 de签名 notes.
Read references/douyin-lea