// the find
openilink/openilink-hub
开源微信 Bot 管理平台 + App 应用市场 | Self-hosted WeChat Bot Platform with App Marketplace | Lark · Slack · Discord · DingTalk · GitHub · Notion · 20+ Apps | AI Tools | 7 Language SDKs
OpeniLink Hub is a self-hosted management platform for WeChat bots built on iLink, WeChat's first official bot protocol launched in March 2026. It wraps the raw iLink protocol into a full platform: multi-account management, an app marketplace with 20+ integrations (Slack, GitHub, Notion, etc.), and multi-channel message dispatch via WebSocket, Webhook, or AI auto-reply. The target audience is developers who want to build on WeChat without handling session management, CDN encryption, and message routing themselves.
The single-binary zero-config deployment is genuinely good — SQLite by default means `docker run` and you're done, no external dependencies, with a clean escape hatch to PostgreSQL via one env var. The mock server for local App development reuses the actual Bot API handlers rather than stubbing them, so what works locally works in production — a detail most projects skip. Message tracing with full delivery timelines per message is built in from the start, not bolted on later, and the store abstraction sharing a test suite across both SQLite and PostgreSQL backends keeps them honest. The App marketplace architecture — PKCE OAuth installation, independent service model, echo template — is a real extension point rather than a config-file plugin system.
The disclaimer in the first line is the most important sentence in the README: this is built on a reverse-engineered protocol, WeChat can break it with a server-side change at any time, and the authors commit to taking it down if pressured. Any production workflow depending on this is betting against WeChat's track record with third-party automation. The 24-hour session window is a hard platform constraint that can't be auto-renewed silently — someone has to send a message to reset it, which makes the 'automated' use case fragile in practice. The SQLite and PostgreSQL migration tracks have diverged badly (38 PG migrations vs 9 SQLite), and there's no documented path from one to the other, so anyone starting on SQLite who outgrows it is on their own. The 'seven language SDKs' claim overstates reality — C# is listed as in development and Lua requires building from source, so the cross-language story is mostly Go and Node.js with stubs elsewhere.