Archestra.AI

Archestra.AI

Centralized MCP platform to manage MCP servers, registry, orchestrator, security, cost monitoring, and observability.

279
Stars
45
Forks
20
Releases

Overview

Open-source, enterprise-ready MCP gateway and management platform designed to simplify and secure enterprise usage of MCPs. It provides a centralized MCP toolbox with observability and control built on a strong security foundation. The platform enables running MCP servers in Kubernetes via a native orchestrator that manages server state, API keys and OAuth. It includes a private MCP registry with governance to share MCPs across teams, supporting both self-hosted and remote, self-built and third-party MCPs. Security sub-agents isolate dangerous tool responses to prevent prompt injections, while dynamic tool call blocking helps prevent data exfiltration. Cost monitoring and limits are available per team, per agent, or per organization, with a dynamic optimizer that can reduce costs by automatically switching to cheaper models. Observability features provide metrics, traces, and logs to analyze per-org, per-agent, and per-team usage and performance. The platform is production-ready, with tooling such as a Terraform provider and Helm chart to support production deployment and quickstart guidance for Docker-based installation.

Details

Owner
archestra-ai
Language
TypeScript
License
GNU Affero General Public License v3.0
Updated
2025-12-07

Features

ChatGPT-like chat with MCPs

Conversational interface to interact with MCPs using a private prompt registry for secure, shareable prompts.

Private MCP registry with governance

Self-hosted and remote MCP registry to share MCPs across teams with governance controls, including self-built and third-party MCPs.

Kubernetes-native MCP orchestrator

Run MCP servers in Kubernetes and manage their state, API keys, and OAuth access.

Security sub-agents

Isolate dangerous tool responses from the main agent to prevent prompt injections.

Dynamic tool call blocking

Block potentially dangerous tool calls to prevent data exfiltration from MCP interactions.

Cost monitoring, limits and dynamic optimization

Per-team/ per-org cost monitoring with dynamic optimization to reduce costs by switching to cheaper models automatically.

Observability

Metrics, traces and logs to analyze per-org, per-agent and per-team usage and performance.

Production-ready tooling

Terraform provider and Helm chart support for production deployment and quickstart guidance.

Audience

Platform teamsCentralize MCP servers, manage access, credentials, costs, and observability across the organization.
DevelopersDeploy MCP servers across the organization and build agents with security considerations handled by the platform.
ManagementEnable enterprise-wide MCP adoption with cost savings and full visibility into usage and data access.

Tags

MCPMCP registryMCP orchestratorKubernetessecurityprompt-injection defensedata exfiltration preventioncost managementobservabilityproduction-readyTerraformHelmself-hostedgovernanceprivate registry