# Agent Infrastructure

> Gateways, routers, registries, and runtimes that agent and MCP stacks run on.

Agent infrastructure is the plumbing beneath agents and MCP stacks — model gateways and routers, tool and server registries, sandboxed runtimes, memory and state stores, and observability — the layer teams reach for once a prototype meets real traffic. Rather than a specific agent, these products solve cross-cutting production problems: routing across model providers for cost and failover, executing untrusted tool calls safely, controlling and auditing what agents can access, and monitoring what they actually did. Choose based on the specific layer you're missing, support for open standards like MCP and OpenAI-compatible APIs so you avoid lock-in, and the production essentials — reliability, security, latency, and cost controls. A model gateway earns its place by cutting spend and adding failover without adding meaningful latency; a sandbox earns its place by containing what a misbehaving agent can do to the rest of your systems. Look at how observability surfaces token cost, tool calls, and failures per run, and whether isolation is strong enough for code you don't fully trust. Offerings range from open-source components you self-host to managed, usage-based services. Adopt this layer when you're running many agents, many users, or untrusted tools and need routing, isolation, cost governance, and monitoring you can't reasonably build and maintain yourself — before an incident forces the decision for you.

23 tools indexed. Sponsored placements are labeled and never ranked; open-source ordering reflects real GitHub adoption. Browse on the web: https://agentsearchengine.app/category/agent-infrastructure

## Open-source index (ranked by adoption)
1. [daytona](https://agentsearchengine.app/agents/daytona) — Daytona is a Secure and Elastic Infrastructure for Running AI-Generated Code · Open source · Infrastructure · 72k★ · active 10 days ago · best for running ai-generated code in secure, elastic sandboxes
2. [mem0](https://agentsearchengine.app/agents/mem0) — Universal memory layer for AI Agents · Open source · Infrastructure · 62k★ · Python · Apache-2.0 · active 2 days ago · best for adding a persistent memory layer to agents across sessions and users
3. [cua](https://agentsearchengine.app/agents/cua) — Open-source infrastructure for Computer-Use Agents. Sandboxes, SDKs, and benchmarks to train and evaluate AI agents that can control full desktops (macOS, Linux, Windows). · Open source · Infrastructure · 21k★ · Python · MIT · active today · best for background computer-use across macos, windows, and linux without stealing focus
4. [gateway](https://agentsearchengine.app/agents/gateway) — A blazing fast AI Gateway with integrated guardrails. Route to 1,600+ LLMs, 50+ AI Guardrails with 1 fast & friendly API. · Open source · Infrastructure · 13k★ · TypeScript · MIT · active 2 months ago · best for routing a single fast api across 250+ llms from 40+ providers
5. [Bindu](https://agentsearchengine.app/agents/bindu) — Bindu: The identity, communication, and payments layer for AI agents. · Open source · Infrastructure · 8.0k★ · Python · active 14 days ago
6. [osaurus](https://agentsearchengine.app/agents/osaurus) — Own your AI. The native macOS harness for AI agents -- any model, persistent memory, autonomous execution, cryptographic identity. Built in Swift. Fully offline. Open source. · Open source · Infrastructure · 7.5k★ · Swift · MIT · active today
7. [steel-browser](https://agentsearchengine.app/agents/steel-browser) — Open Source Browser API for AI Agents & Apps. Steel Browser is a batteries-included browser sandbox that lets you automate the web without worrying about infrastructure. · Open source · Infrastructure · 7.4k★ · TypeScript · Apache-2.0 · active 6 days ago
8. [registry](https://agentsearchengine.app/agents/registry) — A community driven registry service for Model Context Protocol (MCP) servers. · Open source · Infrastructure · 7.1k★ · Go · active 4 days ago
9. [bifrost](https://agentsearchengine.app/agents/bifrost) — Fastest enterprise AI gateway (50x faster than LiteLLM) with adaptive load balancer, cluster mode, guardrails, 1000+ models support & <100 µs overhead at 5k RPS. · Open source · Infrastructure · 7.0k★ · Go · Apache-2.0 · active today
10. [klavis](https://agentsearchengine.app/agents/klavis) — Klavis AI: MCP integration platforms that let AI agents use tools reliably at any scale · Open source · Infrastructure · 5.8k★ · Python · Apache-2.0 · active 2 months ago
11. [lemonade](https://agentsearchengine.app/agents/lemonade) — Lemonade helps users discover and run local AI apps by serving optimized LLMs right from their own GPUs and NPUs. Join our discord: https://discord.gg/5xXzkMu8Zk · Open source · Infrastructure · 5.2k★ · C++ · Apache-2.0 · active today
12. [byterover-cli](https://agentsearchengine.app/agents/byterover-cli) — ByteRover CLI (brv) - The portable memory layer for autonomous coding agents (formerly Cipher) · Open source · Infrastructure · 4.9k★ · TypeScript · active 1 month ago
13. [skillhub](https://agentsearchengine.app/agents/skillhub) — Self-hosted, open-source agent skill registry for enterprises. Publish & version skill packages, govern with RBAC and audit logs, deploy on-premise with Docker or Kubernetes. · Open source · Infrastructure · 4.8k★ · Java · Apache-2.0 · active today
14. [mcpo](https://agentsearchengine.app/agents/mcpo) — A simple, secure MCP-to-OpenAPI proxy server · Open source · Infrastructure · 4.3k★ · Python · MIT · active 2 months ago
15. [archestra](https://agentsearchengine.app/agents/archestra) — Enterprise AI Platform with guardrails, MCP registry, gateway & orchestrator · Open source · Infrastructure · 4.1k★ · TypeScript · active 1 day ago
16. [metamcp](https://agentsearchengine.app/agents/metamcp) — MCP Aggregator, Orchestrator, Middleware, Gateway in one docker · Open source · Infrastructure · 2.6k★ · TypeScript · MIT · active 1 month ago
17. [VCPToolBox](https://agentsearchengine.app/agents/vcptoolbox) — VCP 部署在 AI 模型 API 与前端应用之间，是面向AGI OS开发和探索的工业级基建示范项目。通过统一指令协议、多层级持久化记忆、分布式插件引擎及多 Agent 协作框架，将原本“无状态、无记忆、无工具调用能力”的大语言模型，彻底改造成拥有永久自我意识、物理世界操作权及群体协作智能的完整智能体系统。 · Open source · Infrastructure · 2.2k★ · JavaScript · active today
18. [mcp-router](https://agentsearchengine.app/agents/mcp-router) — A Unified MCP Server Management App (MCP Manager). · Open source · Infrastructure · 2.1k★ · TypeScript · active 1 month ago
19. [toolhive](https://agentsearchengine.app/agents/toolhive) — ToolHive is an enterprise-grade platform for running and managing Model Context Protocol (MCP) servers. · Open source · Infrastructure · 2.0k★ · Go · Apache-2.0 · active 1 day ago
20. [ai-devkit](https://agentsearchengine.app/agents/ai-devkit) — The control plane for AI coding agents. · Open source · Infrastructure · 1.6k★ · TypeScript · active 1 day ago
21. [MCP-Bridge](https://agentsearchengine.app/agents/mcp-bridge) — A middleware to provide an openAI compatible endpoint that can call MCP tools · Open source · Infrastructure · 928★ · Python · MIT · active 7 months ago
22. [agentlytics](https://agentsearchengine.app/agents/agentlytics) — Comprehensive analytics dashboard for AI coding agents — Cursor, Windsurf, Claude Code, VS Code Copilot, Zed, Antigravity, OpenCode, Command Code · Open source · Infrastructure · 554★ · JavaScript · active 1 month ago
23. [mlx-serve](https://agentsearchengine.app/agents/mlx-serve) — Native LLM inference server for Apple Silicon. OpenAI + Anthropic API compatible. No Python. Includes MLX Core macOS app with chat, agent mode, and tool calling. · Open source · Infrastructure · 405★ · Zig · MIT · active today

## Frequently asked
**What is agent infrastructure?**
It's the supporting layer for running agents in production — model gateways/routers, tool and MCP registries, sandboxed runtimes, and observability — rather than the agents themselves.

**When do I need agent infrastructure?**
When prototypes meet real traffic and you need reliable model routing, secure tool execution, cost controls, and monitoring across many agents or users.

**How does this relate to MCP servers?**
MCP servers expose individual tools to agents; infrastructure (gateways, routers, registries) manages, secures, and scales access to many of them.

---
Source: [Agent Search Engine](https://agentsearchengine.app). Full directory in one file: https://agentsearchengine.app/llms-full.txt