MCP Servers
MCP servers are the core of MCPForest. Each server exposes a set of tools that can be accessed by any MCP-compatible client. This guide covers creating, deploying, and managing servers.
What is an MCP Server?
A Model Context Protocol (MCP) server provides a standardized interface for AI models and applications to access external tools and data sources. MCPForest hosts and manages these servers for you — you define the tools and we handle the infrastructure.
Creating a Server
From the dashboard, click "New Server" to start the creation wizard. The flow has four stages:
1. Upload Spec
Provide a server name, upload your OpenAPI specification (.json or .yaml), and optionally set Basic Auth credentials if your API requires authentication.
2. Select Tools
MCPForest parses your spec and generates tool definitions. Browse, search, enable or disable individual tools, and rename them as needed.
3. Build Workflows
Optionally chain tools into workflows using the visual canvas editor. Drag tools, connect nodes, and configure input mappings between steps.
4. Review & Deploy
Review your configuration — server name, enabled tools, workflows, and endpoint URLs. Then click Deploy to go live.
Server Statuses
Each server displays a status badge on the dashboard and detail page:
| Status | Description |
|---|---|
| Running | Server is deployed and accepting requests |
| Deploying | Server is being provisioned |
| Building | Server image is being built |
| Stopped | Server has been manually stopped |
| Failed | Server encountered a deployment or runtime error |
| Draft | Server has been created but not yet deployed |
Server Detail Page
Click on any server from the dashboard to view its detail page. From here you can:
- Deploy / Redeploy — Deploy a draft server or redeploy with updated configuration
- Stop — Stop a running server
- Delete — Permanently remove the server
- View endpoints — See the SSE and Streamable HTTP endpoint URLs
- Integration guides — Get ready-to-copy connection configurations for Claude Desktop, Cursor, Python, Node.js, and cURL
- View tools — See all configured tools with their HTTP methods and paths
- View workflows — See all configured workflows
- Deployment history — View past deployments with status, namespace, pod name, and timestamp
Server Limits
The number of servers you can create depends on your subscription plan. Check the Billing & Plans page for details on plan limits.