docs: add alternative Douyin/XHS MCP option (social-post-extractor-mcp)
Co-authored-by: JNHFlow21 <JNHFlow21@users.noreply.github.com>
This commit is contained in:
@@ -294,6 +294,33 @@ Agent Reach integrates with xiaohongshu-mcp (runs in Docker). After setup, use `
|
||||
Install douyin-mcp-server, then your agent can use `mcporter call 'douyin.parse_douyin_video_info(share_link: "share_url")'` to parse video info and get watermark-free download links. No login required — just share the Douyin link. See https://github.com/yzfly/douyin-mcp-server
|
||||
</details>
|
||||
|
||||
<details>
|
||||
<summary><strong>How to extract scripts from both Douyin and XiaoHongShu with one MCP?</strong></summary>
|
||||
|
||||
If you want one MCP server that can handle:
|
||||
|
||||
- Douyin videos
|
||||
- XiaoHongShu video notes
|
||||
- XiaoHongShu image notes
|
||||
|
||||
and directly write `script.md` + `info.json`, you can point the existing `douyin` mcporter alias at:
|
||||
|
||||
- https://github.com/JNHFlow21/social-post-extractor-mcp
|
||||
|
||||
It keeps backward compatibility with:
|
||||
|
||||
- `parse_douyin_video_info`
|
||||
- `get_douyin_download_link`
|
||||
- `extract_douyin_text`
|
||||
|
||||
and adds unified tools:
|
||||
|
||||
- `parse_social_post_info`
|
||||
- `extract_social_post_script`
|
||||
|
||||
This is useful when your agent workflow is “paste a link, get a script file”.
|
||||
</details>
|
||||
|
||||
---
|
||||
|
||||
## Credits
|
||||
|
||||
@@ -253,6 +253,45 @@ mcporter config add douyin http://localhost:18070/mcp
|
||||
>
|
||||
> 详见 https://github.com/yzfly/douyin-mcp-server
|
||||
|
||||
**可选实现:Douyin + XiaoHongShu unified extractor**
|
||||
> "如果你想把抖音和小红书统一成一个 MCP,并直接输出 `script.md` 和 `info.json`,可以改用 social-post-extractor-mcp。"
|
||||
|
||||
适用场景:
|
||||
|
||||
- 抖音视频转文字稿
|
||||
- 小红书视频笔记转文字稿
|
||||
- 小红书图文笔记正文 + 图片文字提取
|
||||
|
||||
兼容性:
|
||||
|
||||
- 仍然可以注册成 `douyin` 这个 mcporter server 名称
|
||||
- 兼容旧工具名 `parse_douyin_video_info` / `get_douyin_download_link` / `extract_douyin_text`
|
||||
- 同时新增 `parse_social_post_info` / `extract_social_post_script`
|
||||
|
||||
示例配置:
|
||||
|
||||
```bash
|
||||
git clone https://github.com/JNHFlow21/social-post-extractor-mcp.git
|
||||
cd social-post-extractor-mcp
|
||||
uv sync
|
||||
|
||||
mcporter config add douyin \
|
||||
--command /bin/zsh \
|
||||
--arg -lc \
|
||||
--arg "cd '$PWD' && exec '.venv/bin/python' -m social_post_extractor_mcp" \
|
||||
--env ASR_PROVIDER=bailian \
|
||||
--env ASR_MODEL=paraformer-v2 \
|
||||
--env VISION_PROVIDER=bailian \
|
||||
--env VISION_MODEL=qwen3-vl-flash \
|
||||
--env CLEAN_PROVIDER=bailian \
|
||||
--env CLEAN_MODEL=qwen-flash \
|
||||
--env BAILIAN_API_KEY=YOUR_BAILIAN_API_KEY
|
||||
```
|
||||
|
||||
> 这个实现更适合“把链接直接交给 Agent,然后拿到脚本文件”的工作流。
|
||||
>
|
||||
> 详见 https://github.com/JNHFlow21/social-post-extractor-mcp
|
||||
|
||||
**LinkedIn (可选 — linkedin-scraper-mcp):**
|
||||
> "LinkedIn 基本内容可通过 Jina Reader 读取。完整功能(Profile 详情、职位搜索)需要 linkedin-scraper-mcp。"
|
||||
|
||||
|
||||
Reference in New Issue
Block a user