Uber's Model Context Protocol Architecture: MCP Gateway at Enterprise Scale

X postsantiResource

Santi reviews an architectural deep-dive from Uber Engineering on scaling MCP infrastructure. Managing over 800 MCP servers and 5,000 tools, Uber relies on a centralized MCP Gateway that handles protocol translation, automated tool generation with AutoCrawler, and dynamic discovery through Omni MCP.

Key points

  1. Centralized MCP Gateway: Acts as a proxy layer between AI agents and internal backend systems, translating MCP calls into HTTP, gRPC, or TChannel.
  2. Automated API Discovery: AutoCrawler uses Cadence workflows to scan IDL registries (Protobuf, Thrift), translating schemas and enriching tool descriptions with LLMs.
  3. Dynamic Discovery via Omni MCP: Avoids blowing prompt context limits by having agents dynamically discover servers and load specific tool schemas on demand.
  4. Governance by Default: Discovered tools are placed in the MCP Registry in a disabled-by-default state to enforce security and access control.

Why it matters

Widespread enterprise adoption of AI agents quickly leads to tooling fragmentation and exhausted context windows. Uber provides a blueprint for adapting the traditional API gateway pattern to agentic tool ecosystems.

Analysis

Uber's approach highlights the shift from isolated agent proofs of concept to enterprise platform engineering. Rather than requiring service teams to manually write MCP servers, the gateway standardizes protocols and decouples agent interfaces from underlying transport layers. The combination of automated schema crawling and just-in-time schema loading (Omni MCP) resolves both operational maintenance and context token constraints.

Original post

santi

@santtiagom_ · 3 October 2026

Uber publicó cómo armó su infraestructura de MCP y hay varias ideas muy buenas acá. imaginate una empresa con miles de servicios internos y muchos equipos conectando agentes a esos sistemas por su cuenta. cuando eso empieza a crecer, terminás con tooling fragmentado, infra t.co/DccvOOYhos t.co/RLtEvs43mK

6.6k likes89 replies11.9k bookmarks699.7k views

View on X
Selected replies (5)
  • @MatiasMatthews @santtiagom_ A esa escala, el catálogo que ve cada agente también importa: descubrir una herramienta no debería darle permiso para usarla. Filtrarlo según quién actúa evita ofrecerle caminos que después van a terminar bloqueados.
  • @iLegend_AI @santtiagom_ 工具一多，模型选错的概率不是线性涨的。几千个团队各包一层同一个内部服务，最后就是一打名字几乎一样的工具，挑哪个全看描述写得怎么样。比网关更该先做的是工具注册表：一个能力一个规范名，别名全映射过去。不然代理的失败会伪装成「模型不行」。
  • @juanmacias @TheDevilOps @santtiagom_ Algo tengo, por ahora 5 cli y uno de ellos sustituye a 10 MCPs Si buscas info verás que el uso de tokens es mucho menor
  • @juanmacias @ajimenezsa @santtiagom_ Depende del MCP. Si maneja archivos grandes, o un volumen alto, el cli va a ir mejor. Adicionalmente, no tienes que reiniciar cuando hay un cambio. Ahora mismo cuando cambia un MCP tengo que reiniciar Claude Code
  • @chooserich @santtiagom_ This is definitely big company problems 😂 The irony is this harkens back to service oriented architecture but even less efficient. Reaffirms my belief that deterministic services are still better for most things (at least for the time being)

Mentioned links

Summary generated by AI (Gemini 3.8 Flash, 5 October 2026). It may contain errors – the original source is authoritative.

Related entries

Similar in content, found by the AI search.

  • Link:Rhys Sullivan

    Executor: Universal MCP Gateway for AI Agents

    Executor connects OpenAPI specs, GraphQL endpoints, and MCP servers into a single MCP gateway. Instead of manually configuring credentials and tools across multiple agent clients, it exposes a single execution tool that retrieves integrations on demand inside a sandboxed environment.

    AI & LLMs· Tool

  • Repository:xdevplatform/xmcp

    xmcp: FastMCP Server for the X API

    xmcp is a Python-based Model Context Protocol (MCP) server that exposes the X API OpenAPI specification as callable tools using FastMCP, allowing AI agents to interact with X endpoints.

    853starsPython

    AI & LLMs· Tool

  • Link:Arcade

    Arcade: Enterprise Actions Runtime and Governance Layer for AI Agents

    Arcade is an actions runtime designed to govern and secure enterprise AI agents interfacing with internal and external tools. It acts as an enforcement point between agents and connected systems, handling identity provider authentication, scoped authorization, and audit logging, with deep support for the Model Context Protocol (MCP).

    AI & LLMs· Tool

  • Repository:FeiZhuLulu/Agent-Bridge

    Agent Bridge: Orchestrating Local Coding Agents via MCP

    Agent Bridge is an open-source Python tool connecting local coding agents using the Model Context Protocol (MCP). It allows users to interact with a single coordinator agent that delegates tasks to multiple specialized CLI worker agents.

    31starsPython

    AI & LLMs· Tool

Back to the knowledge base

Let's talk about your project

Get in touch

Locations

  • Mattersburg
    Johann Nepomuk Bergerstraße 7/2/14
    7210 Mattersburg, Austria
  • Vienna
    Ungargasse 64-66/3/404
    1030 Wien, Austria

Parts of this content were created with the assistance of AI.