Skip to content
BSWEN
Blog

How Agent Reach Supports Chinese Internet Platforms for AI Agents (Bilibili, XiaoHongShu, Xueqiu, V2EX)

Cowrie
Cowrie
Dev @ Bswen

Problem

A significant portion of high-quality technical content lives on Chinese platforms — Bilibili video tutorials, XiaoHongShu product reviews, Xueqiu investment discussions, V2EX tech community. Most general-purpose AI agent tools ignore these platforms entirely.

When I tried to make my AI agent read Bilibili subtitles or search XiaoHongShu, none of the standard tools worked.

Chinese Platform Support

Agent Reach provides dedicated support for 5 Chinese platforms, each with its own channel file.

Bilibili (B站)

agent_reach/channels/bilibili.py
backends = ["bili-cli", "OpenCLI", "B站搜索 API"]

yt-dlp was removed from this channel in June 2026 — Bilibili’s risk control 412-blocked it. bili-cli works without login for search, hot videos, rankings, and video detail.

Bilibili commands
bili search "AI tutorial" --type video
bili hot
bili info BV1GJ411m8R8

XiaoHongShu (小红书)

agent_reach/channels/xiaohongshu.py
backends = ["OpenCLI", "xiaohongshu-mcp", "xhs-cli"]

On desktop, OpenCLI reuses your Chrome browser session — browse XHS in Chrome, agent reads it directly. On servers, xiaohongshu-mcp uses a headless browser with QR code login.

The channel includes format_xhs_result() and _clean_note() functions that strip structural redundancy from API responses, specifically to reduce LLM token usage.

XiaoHongShu via OpenCLI on desktop
opencli xiaohongshu search "product review"
opencli xiaohongshu note URL
XiaoHongShu via MCP on server
mcporter call 'xiaohongshu.search_feeds(keyword: "手机")' --timeout 120000

Xueqiu (雪球)

Stock quotes, stock search, hot posts, and hot stock rankings. Uses Chrome/Chromium’s cookie store for auth token extraction. Extensive fixes in v1.3.1 for cookie loading strategy, User-Agent, Referer headers, and endpoint migration.

Configure cookies for Xueqiu
agent-reach configure --from-browser chrome

V2EX

Hot topics, node topics, topic detail with replies, and user profiles. Uses the public JSON API — zero config required. Added in v1.3.0.

V2EX hot topics
curl -s "https://www.v2ex.com/api/v2/topics/hot"

Xiaoyuzhou Podcast (小宇宙)

Podcast audio transcription via Groq Whisper. Requires a free Groq API key.

Transcribe a podcast episode
bash ~/.agent-reach/tools/xiaoyuzhou/transcribe.sh https://www.xiaoyuzhoufm.com/episode/xxxxx

Desktop vs Server Split

Agent Reach handles the unique challenge of Chinese platforms by being environment-aware:

  • Desktop — OpenCLI reuses browser login sessions (zero friction)
  • Server — Headless browsers (xiaohongshu-mcp) with QR code login provide the same capability

Anti-Scraping Resilience

Chinese platforms have aggressive anti-scraping measures. Agent Reach’s multi-backend architecture is critical here:

  • Bilibili 412-blocked yt-dlp → seamless switch to bili-cli
  • Xueqiu’s xq_a_token is dynamically generated by frontend JavaScript → browser-level cookie extraction required

The entire README and install guide are available in Chinese. English, Japanese, and Korean translations are in docs/.

Summary

In this post, I showed how Agent Reach supports 5 Chinese platforms that most AI agent tools ignore. The key point is that each platform has a dedicated channel with environment-aware backends — desktop reuses your browser session, servers use headless browsers. The multi-backend routing handles anti-scraping changes transparently.

Final Words + More Resources

My intention with this article was to help others share my knowledge and experience. If you want to contact me, you can contact by email: Email me

Here are also the most important links from this article along with some further resources that will help you in this scope:

Oh, and if you found these resources useful, don’t forget to support me by starring the repo on GitHub!

Comments

Tags: