Mining WeChat for Commitments, Leads, and Loose Ends

What it does

wechat-intelligence-hub is a local-first intelligence layer that reads your authorized WeChat database—without hooking, injecting, or re-signing the app—and turns chat history into searchable, actionable output. It generates daily briefings, flags pending replies and unfulfilled promises, surfaces business opportunities, and tracks reconnection leads. Everything runs against your local machine; the system is read-only and never sends messages.

The interesting bit

The project treats chat not as ephemeral conversation but as a structured intelligence feed. It wraps a clean-room WeChat reader core inside Codex skills, so you can ask natural-language questions like what you promised last week or which contacts need follow-up. Reports come in both Markdown and an interactive HTML dashboard with search, filters, and dark mode.

Key highlights

• Read-only reader core explicitly avoids key extraction, injection, or UI automation
• Natural-language interface via Codex ($wechat-intelligence-hub) with profile-based personalization for prioritization
• Generates time-range reports covering action items, group chats, and contact priorities
• Dual output: archival Markdown plus an interactive HTML report with global search and topic radar
• Falls back to macOS notification previews for inbound messages when full database access is unavailable

Caveats

• Still at v0.9.2-preview.2; database compatibility may shift with WeChat updates
• Full history requires explicit local database authorization and access materials—there is no automatic, unattended decryption
• macOS new-machine access paths remain unverified, and notification-preview mode only covers inbound traffic, not complete chat logs

Verdict

Worth a look if you run business or project coordination through WeChat and want to stop manually trawling group chats for action items. Skip it if you expect a one-click install or are uncomfortable authorizing local database access.