ReadGZH

ReadGZH is an online service that converts WeChat official account articles into AI-readable formats. WeChat official accounts tightly restrict external requests — ChatGPT, Claude, and DeepSeek requesting mp.weixin.qq.com directly often can't get the article body — so it fetches and cleans via a server-side proxy, outputting pure text with zero JS dependencies, built for researchers and developers who need to feed official account long-forms to AI.

Core capabilities

• MCP Server: built-in remote MCP with 5 tools (read/list/search, etc.), letting agents fetch article bodies with zero installation
• REST API: GET /rd?url=<article link> returns the body directly, with format=text for Markdown and part=N chunking
• Server-side cleaning: strips inline style/class/data-* attributes, removes WeChat-private tags, and collapses empty tags, outputting clean SSR HTML
• Shared cache: cache hits read at 0 credits, with long articles auto-chunked at roughly 40KB
• Image proxy: forwards WeChat images through a CDN, easing expired original-image links
• AI summaries: Pro users can generate structured summaries with mode=summary

Use cases Feeding official account long-forms to ChatGPT/Claude/DeepSeek for summarization, translation, and comparison; developers building "WeChat article → Markdown" data pipelines for agents/RAG; enterprises needing shared quotas and invoices for content analysis.

Unique advantages Where Jina Reader frequently returns 403 on WeChat official accounts, ReadGZH specializes vertically in anti-scraping penetration; versus open-source solutions requiring local deployment, it is cloud-hosted, zero-install infrastructure with shared caching and image proxying.

Editorial verdict Both the API and MCP tested live in this review; the free 30 credits daily is a friendly threshold and Lite starts at ¥9/month — few one-stop alternatives exist for this "WeChat content → AI" need. Caveats: the operator is an anonymous indie developer with no company entity or ICP filing; the public "article library" indexing others' articles raises copyright disputes (the site offers a takedown channel and disclaimers); key metrics are self-reported with no third-party reviews. Respect original authors' copyrights and don't process private content.