wechat-article-hub

Self-hosted hybrid for downloading WeChat Official Account articles in batch, with stealth scraping and residential-proxy rotation.

Combines:

• UI & batch flow — vendored fork of wechat-article-exporter
• Stealth content fetch — vendored fork of bzd6661/wechat-article-for-ai, powered by Camoufox (stealth Firefox)
• Residential proxy rotation — configurable via config.yaml

Quickstart

git clone <repo> wechat-article-hub && cd wechat-article-hub
cp .env.example .env
# edit .env: set WECHAT_RESIDENTIAL_PROXY
docker compose up
# open http://localhost:3000

Log into 微信公众平台 via QR. Select公众号, filter articles, export to HTML / Markdown / Excel / DOCX / JSON / TXT.

Architecture

Two services in compose:

• nuxt (port 3000) — Nuxt 3 SPA + Nitro server. Hosts the UI and proxies authenticated MP requests with the user's cookies.
• scraper (internal port 8000) — FastAPI + Camoufox pool. Owns the residential proxy and per-article fetch.

Nuxt → scraper communication is HTTP-only via NUXT_SCRAPER_URL. The two services have no shared code; you can deploy them anywhere as long as Nuxt can reach the scraper.

See docs/superpowers/specs/2026-04-19-wechat-article-hub-design.md for the full design.

Status

v0.1.0 — verified end-to-end against 微信公众平台 on 2026-04-20. HTML and Markdown export both work. See CHANGELOG.md for details, known limitations, and upstream credits.

Configuration

All behavior is in config.yaml (mounted into both containers). All secrets / per-deploy infra is in .env. See docs/deployment-personal.md and docs/deployment-enterprise.md for examples.

Documentation

• CHANGELOG
• Personal deployment
• Enterprise deployment
• Release smoke checklist
• Design spec

License

MIT (this repo). See upstream projects for their respective licenses.