> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.beatapi.io/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.beatapi.io/_mcp/server.

# Use the Coze HTTP node to find Xiaohongshu notes or read a WeChat article

> Configure Coze authentication, pass a keyword or URL, and return evidence to the next node.

Add one **HTTP Request** node in Coze that POSTs to `/v1/social-data/call`. For editorial leads, pass a keyword such as “camping coffee” to `xiaohongshu.app_v2.search_notes` and return Xiaohongshu candidates. For a WeChat tutorial an editor already selected, pass the article URL to `wechat_mp.v2.fetch_article_detail_h5` and return the article text and source. Build the branch you need first.

1. In **Start**, define a string `keyword` for the notes workflow and a string `article_url` for the WeChat workflow. Use a verified article link for `article_url`. Build separate workflows, or branch by source.
2. Add an **HTTP Request** node. Set the API URL to `https://api.beatapi.io/v1/social-data/call` and choose POST. Select **Bearer Token** authentication and supply the BeatAPI key through protected configuration, so the request headers do not repeat Bearer. Choose JSON for the body and insert the string values with the editor's variable picker.

```json
{"action":"xiaohongshu.app_v2.search_notes","params":{"keyword":"<keyword>","page":1}}
```

```json
{"action":"wechat_mp.v2.fetch_article_detail_h5","params":{"url":"<article_url>"}}
```

3. Pass the HTTP status and `body` to the validation step. Parse `body` and check `status`, `error` and `request_id`. For notes, keep the source identifiers and available titles from `data.data.items[].note`. For WeChat, take the original link, title and `content_text` from `data.content`. Mark a missing `content_text` for follow-up, and keep `desc` in a separate summary field.
4. In **End**, return the result and the request ID as variables. Test a successful result, an empty list and a failed call before giving the workflow to an agent. Review [BeatAPI pricing](https://beatapi.io/pricing) and the execution history before turning on node retries.

If one bot offers both tasks, add a task type to the Start node and branch on the task type: run note search for a keyword and article lookup for a known URL. If the branch's required input is missing, return “Please provide a keyword” or “Please provide a WeChat article URL” without making a request.

## Check one result in the End node

Suppose an editor supplies a WeChat tutorial. Map the verified URL, the returned title and `content_text` to separate End-node output fields. If `content_text` is empty, set `status` to `body_unavailable` for follow-up. When the response contains `desc`, pass `desc` to the editor as `summary`. The JSON below shows example End-node fields.

```json
{
  "status": "body_unavailable",
  "source_url": "https://mp.weixin.qq.com/s/example",
  "title": "Camping coffee equipment checklist",
  "body_text": null,
  "summary": "A list of beginner gear and packing suggestions.",
  "request_id": "example-request-id"
}
```

For the Xiaohongshu branch, return a candidate array with note IDs and available titles, plus the request ID of the search. For a checklist with citations, have an editor select notes before fetching note details. After one successful run, test an empty search and check that the editor receives an empty candidate list rather than invented material.

To start from an account, use the [WeChat article list](/guides/wechat-incremental-articles), select an article and pass the article URL to the detail node. Keep each input, status and request ID together, so a downstream node can tell evidence from an error.

[Coze HTTP Request node](https://docs.coze.cn/guides_http_node)