Summary

This project provides a local-first "skill" for AI assistants (Codex/Claude Code) to manage and analyze scattered WeChat official account content. It enables users to build a persistent, searchable local repository of articles, comments, and interaction data, transforming ephemeral content into a reusable knowledge base for creators, researchers, and knowledge managers.

How It Works

The core approach is a local-first skill emphasizing data privacy. It offers four scenarios: syncing subscribed accounts, researching specific accounts' content and reader feedback, on-the-fly saving via a WeChat collection session, and archiving scattered article links. The system uses SQLite for local state management, outputting content as Markdown, images, and an index CSV for subsequent AI analysis.

Quick Start & Requirements

Installation involves cloning the repository into the Codex or Claude Code skills directory (~/.codex/skills/ or ~/.claude/skills/). Specific features require: WeChat Official Account backend login (history export), macOS WeChat client with mitmproxy (collection/proxy history), and macOS system proxy management. See SKILL.md and references for details.

Highlighted Details

• Four Core Scenarios: Supports syncing updates, detailed account research (topic, structure, feedback), interactive WeChat collection, and batch link archiving.
• Rich Data Capture: Captures selected comments and interaction data (views, likes, comments) alongside article text.
• Local Repository: Organizes content into ~/Downloads/wechat-articles/<公众号名>/ with Markdown articles, images, and an SQLite index.
• Privacy-Focused: Operates locally without third-party services.

Maintenance & Community

The provided documentation does not detail specific contributors, sponsorships, or community channels.

Licensing & Compatibility

The license type is not specified. The tool is a skill for Codex/Claude Code with specific OS dependencies (macOS) for certain features. It is not a SaaS product.

Limitations & Caveats

Processes only public articles, not bypassing paywalls or private content. Captures comments/interaction data as returned by the interface, not guaranteeing complete threads. Certain features are macOS-specific.