archestra-ai/archestra

Enterprise AI Platform with guardrails, MCP registry, gateway & orchestrator

What it solves

Archestra provides a unified, secure enterprise platform for deploying and managing AI agents and LLM integrations. It solves the fragmentation and security risks associated with running multiple AI tools in a corporate environment by centralizing access, governance, and observability.

How it works

The platform acts as a comprehensive gateway and runtime environment that integrates several core components:

  • LLM & MCP Gateways: Manages connections to various providers (Anthropic, OpenAI, etc.) with cost limits, virtual API keys, and dynamic routing. It also handles Model Context Protocol (MCP) authentication using OAuth.
  • Agent Runtime: Executes agents with support for scheduled, email, or webhook triggers, sub-agent delegation, and sandboxed code execution.
  • Knowledge Base: Implements RAG (Retrieval-Augmented Generation) via connectors to existing data stacks.
  • Governance & Security: Enforces RBAC, SSO, and deterministic guardrails (such as Dual-LLM verification) to prevent unsafe tool calls.
  • Infrastructure: Built for Kubernetes with a dedicated operator and OpenTelemetry/Prometheus for observability.

Who it’s for

It is designed for enterprises and development teams that need to deploy production-ready AI agents and LLM applications while maintaining strict security, cost control, and observability standards.

Highlights

  • Enterprise-grade security: Includes SSO, RBAC, and sandboxed code execution.
  • Unified Gateway: Single entry point for users and developers to access multiple LLM providers and MCP tools.
  • Agent Orchestration: Supports complex agent workflows including sub-agent delegation and a private MCP registry.
  • Integrated Observability: Built-in OpenTelemetry traces and Prometheus metrics for monitoring and cost tracking.

Related

  • Project
    onyx-dot-app/onyx-foss

    Onyx is an MIT‑licensed, self‑hosted platform that indexes data from 50+ workplace apps, builds a hybrid vector/keyword store, and exposes it to LLMs via an API and chat UI. It provides agentic RAG, custom AI agents, web search, sandboxed code execution, and integrations (Slack, Discord, Chrome, embeddable widget). Deployable via Docker, Kubernetes, or Helm, it offers a full‑featured “Standard” mode and a lightweight “Lite” mode. Designed for secure, private AI assistants, it targets teams that need internal knowledge retrieval and action‑taking capabilities.

  • Project
    Team-Commonly/commonly

    An open-source coordination workspace where humans and AI agents collaborate using shared pods, persistent memory, and a GitHub-synced task board.

  • Project
    mezmo/aura

    AURA is an open‑source SRE‑agent platform that lets you run AI‑driven incident‑investigation assistants inside your own infrastructure. It supports multiple LLM providers, configurable guardrails, OpenTelemetry tracing, and dozens of integrations (Kubernetes, Prometheus, PagerDuty, etc.). Deployable as a CLI, Docker image, Helm chart, or embedded Rust library.

  • Project
    rossoctl/rossoctl

    Rossoctl is an open‑source, Kubernetes‑native platform that intercepts AI agents via the RossoCortex data‑plane, enforcing identity, authorization, resilience and observability. It adds reusable services (skills, tools, memory, knowledge base, sandboxes) and admin tooling, enabling trustworthy, production‑grade agent deployments across frameworks.

  • Project
    CyberStrikeus/CyberStrike

    CyberStrike is an open‑source AI agent that turns any LLM (Claude, GPT, Gemini, etc.) into an autonomous penetration‑testing platform. It adds a security‑focused “intelligence layer”, ships 13+ specialised agents and 7,600+ signed security‑skill files, supports 150+ AI providers (including offline models), and can orchestrate real tools locally or on remote Bolt servers. A TUI, optional web UI, and built‑in HackBrowser complete the workflow.