Design and Implementation of Baidu Mini Program Open‑Source Ecosystem and Distribution Assurance

The article introduces the Baidu Mini Program open‑source ecosystem, describing its vision of an open, collaborative platform that serves three types of participants: alliance partners, developers, and end users.

It outlines the four‑layer business architecture of the ecosystem: the Alliance Host layer (e.g., OPPO, VIVO, Xiaomi browsers), the External Alliance Enablement layer (providing onboarding, configuration, integration, and resource services), the Intermediate Service layer (managing hosts, capabilities, distribution, orders, and data), and the Foundation layer (including the Mini Program framework, data services, package services, and security hardening).

The article then discusses the challenges of distributing Mini Programs through partner browsers, focusing on ensuring distribution quality and establishing a full‑process control path from the host browser to Baidu Search.

Technical construction is divided into three main flows: host‑side and Mini Program‑side process integration, capability grading, and distribution flow. Hosts integrate the open‑source SDK, pass CTS (Compatibility Test Suite) testing, and publish their apps. Developers release Mini Program packages, which are automatically subscribed to, scanned by an AST‑based capability scanner, and stored for further analysis.

Capability detection includes host CTS testing, report uploading, and automated parsing to generate a host capability list and grade. Mini Program capability scanning uses an AST tool to extract used APIs, parameters, return values, and key execution paths, followed by capability grading.

Distribution matching uses a level‑based mechanism (L1–L4) to align host capabilities with Mini Program requirements, ensuring only compatible programs are delivered. An additional intervention layer allows administrators to enforce blacklists, private programs, or manual overrides.

The article concludes with reflections on service‑oriented architecture design, emphasizing data centralization, a self‑developed RPC framework for microservice governance, single‑responsibility decomposition, and comprehensive logging and monitoring to support stability and troubleshooting.