Skip to main content
NimbleTools Core is an open-source runtime that enables you to deploy Model Context Protocol (MCP) tools as scalable, production-ready HTTP services in under 60 seconds.

What is NimbleTools Runtime?

NimbleTools Runtime provides a Kubernetes-native platform for deploying and managing MCP services. It automatically handles scaling, networking, and service discovery, allowing you to focus on building great AI tools.

Universal Deployment

Deploy any MCP tool - Python, Go, stdio, or HTTP

Auto-Scaling

Automatically scale replicas based on demand

Production Ready

Built for production workloads

Production Ready

Built-in security, monitoring, and multi-tenancy

Key Features

60-Second Deployment

Deploy command-line MCP tools as HTTP services with a single command:
ntcli server deploy my-mcp-tool

Kubernetes-Native

Built on Kubernetes with custom resource definitions (CRDs) for MCP services:
  • MCPService CRD: Define and manage MCP services declaratively
  • Operator Pattern: Automated reconciliation and health management
  • Service Registry: Built-in discovery and routing

Automatic Scaling

Automatically scale services based on demand:
  • Horizontal pod autoscaling based on CPU/memory
  • Configure min and max replicas
  • Scale up during high load, scale down when idle

Multi-Workspace Support

Isolate services and resources by workspace:
  • Separate namespaces per workspace
  • Resource quotas and limits
  • Team-based access control

Architecture

┌─────────────────────────────────────────────────┐
│                   ntcli / API                    │
└───────────────┬─────────────────────────────────┘

┌───────────────▼─────────────────────────────────┐
│         NimbleTools Operator                     │
│  - MCPService CRD Controller                     │
│  - Auto-scaling Manager                          │
│  - Service Registry                              │
└───────────────┬─────────────────────────────────┘

┌───────────────▼─────────────────────────────────┐
│            Kubernetes Cluster                    │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐      │
│  │   Pod    │  │   Pod    │  │   Pod    │      │
│  │ MCP Tool │  │ MCP Tool │  │ MCP Tool │      │
│  └──────────┘  └──────────┘  └──────────┘      │
└─────────────────────────────────────────────────┘

Use Cases

Deploy internal MCP tools once and make them available to all team members through Claude Desktop, Claude Code, or other MCP clients.
Run MCP tools in production with automatic scaling, monitoring, and high availability.
Test and develop MCP services locally with k3d before deploying to production.
Run your own MCP platform on-premises or in your cloud account with full control.

Runtime vs Studio vs CLI

FeatureRuntimeStudioCLI
DeploymentSelf-hostedCloud-hostedLocal client
Open Source✅ Yes❌ No✅ Yes
UI❌ No✅ Yes❌ No
Auto-scaling✅ Yes✅ Yes❌ N/A
KubernetesRequiredManagedNot required
Best ForSelf-hostingQuick startDevelopment
Studio uses Runtime under the hood: NimbleBrain Studio is powered by NimbleTools Runtime, providing a managed experience with UI, playbooks, and additional features.

Getting Started

Ready to deploy your first MCP service?

Community & Support

License

NimbleTools Runtime is open-source software licensed under Apache 2.0.
I