DecoCMS

•
337
•

Define and compose secure Model Context Protocol (MCP) servers to generate and deploy AI workflows and agents with integrated governance.

DecoCMS is an open-source Context Management System, acting as an "MCP Mesh" for AI, designed to centralize company context, connections, tools, and policies in a secure environment. It empowers teams to build and deploy full-stack AI applications—including agents, workflows, and user interfaces—entirely in TypeScript. DecoCMS provides built-in governance, comprehensive observability, and cost control, facilitating deployment across various environments like Cloudflare Workers, AWS, or local Bun/Deno runtimes, and effectively bridging the gap between AI prototypes and production-ready systems.

主要功能

01253 GitHub stars
02Flexible deployment options, including Cloudflare Workers, AWS, or self-hosting with Bun/Deno runtimes.
03Integrated governance capabilities, offering RBAC, audit trails, and spend caps from initial deployment.
04Unified full-stack TypeScript environment where agents, workflows, and user interfaces share the same repository and types.
05Unified observability providing end-to-end tracing from UI interactions through agent calls to model responses.
06Declarative API (defineTool) for creating type-safe, auditable, and observable Model Context Protocol (MCP) tools.
07MCP-native architecture for composing Model Context Protocol servers with built-in policy, authentication, and observability.

使用案例

01Building and deploying production-scale AI applications that include agents, durable workflows, and custom user interfaces.
02Centralizing and securing company-wide context, tools, and data access for AI agents in a private MCP mesh.
03Exposing governed virtual MCPs (AI Apps) to any client, complete with access control, monitoring, and cost management.