Getting Started with FloMCP
Generate your first production-ready MCP server in under 5 minutes. No manual boilerplate, no configuration headaches.
On This Page
1. What is MCP?
The Model Context Protocol (MCP) is an open standard created by Anthropic that allows AI assistants (Claude, GitHub Copilot, Cursor, etc.) to securely connect to external tools, data sources, and services.
An MCP server exposes tools that the AI can call — for example, reading files, querying a database, calling an API, or running a search. Without MCP, every AI integration requires custom, often insecure glue code. With MCP, you write the server once and use it everywhere.
FloMCP generates complete MCP servers for you — including tool schemas, error handling, input validation, and setup documentation — in under 5 minutes via an AI-assisted 5-step wizard.
2. Create Your Account
- 1Go to floMCP.com/auth/signup and enter your email address.
- 2Check your inbox for a confirmation email and click the link (check spam if you don't see it within 2 minutes).
- 3You're taken to your dashboard. The free plan includes 3 credits on signup — they never expire.
3. The 5-Step Generator Wizard
Navigate to Dashboard → Generate and follow the 5-step wizard:
Step 1 — Describe
Give your server a name and describe what it should do in plain English. Use the built-in example prompts (GitHub, Weather, Database, Slack…) or load a saved prompt. A clear, detailed description yields better tool suggestions and higher-quality generated code.
Step 2 — API Setup (optional)
Configure an external API your server will call: base URL, authentication type (None, API Key, Bearer Token, or OAuth 2.0), API documentation URL (FloMCP fetches and uses it as context), input schema fields, and config options. Skip this step if your server does not call an external API.
Step 3 — Tools
Define the tools your server exposes (up to 20). Three ways to add tools: (1) Generate Tools — Claude suggests 2 tools based on your description; (2) Import Schema — paste OpenAI function JSON, MCP inputSchema, or FloMCP native format; (3) Add Custom Tool — define name, description, parameters, and an optional example output. FloMCP shows a live credit cost estimate as you build.
Step 4 — Resources & Prompts (optional)
Attach static context (Resources) or reusable message templates (Prompts) to your server. Drag & drop .txt, .md, or .json files, or type content directly. Each resource/prompt can be scoped to all tools or a specific tool. This step is optional — skip it if your server needs no injected context.
Step 5 — Review & Generate
Full summary of your configuration with live credit cost. Click Generate — watch the 3-pass engine stream in real time: Pass 1 (schema contract) → Pass 2 (TypeScript implementation) → Pass 3 (quality checklist) → MCP protocol compliance checks → 22 OWASP security checks. When complete, download the ZIP and view your security score.
4. Download & Setup
After generation, click Download ZIP. The archive includes:
- index.ts — Main MCP server entry point
- tools/ — Individual tool handler files
- .env.example — All required environment variables
- README.md — Auto-generated setup instructions specific to your server
- package.json — Dependencies with exact version pins
Install Dependencies
cd your-mcp-server npm install cp .env.example .env # Fill in your API keys in .env npm run build
5. Connect to Claude Desktop
Claude Desktop uses a config file to register MCP servers. Open (or create) the file at:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json
Add your server to the mcpServers object:
{
"mcpServers": {
"my-server": {
"command": "node",
"args": ["/absolute/path/to/your-server/dist/index.js"],
"env": {
"API_KEY": "your-api-key-here"
}
}
}
}Restart Claude Desktop. Your MCP tools will appear in the tool selector (hammer icon) in any conversation.
6. Connect to GitHub Copilot
GitHub Copilot (in VS Code) supports MCP servers via the .vscode/mcp.json workspace config:
// .vscode/mcp.json
{
"servers": {
"my-server": {
"type": "stdio",
"command": "node",
"args": ["/absolute/path/to/your-server/dist/index.js"],
"env": {
"API_KEY": "${input:apiKey}"
}
}
}
}VS Code will prompt for any ${input:varName} values the first time you use the server. See the VS Code MCP docs for full details.
8. Credits & Pricing
FloMCP uses a credit system. The cost of each generation depends on its complexity:
Simple — 1 credit
≤3 tools · content ≤2,000 chars
Complex — 2 credits
4–10 tools · OR content 2,001–3,000 chars
Premium — 3 credits
11–20 tools · OR content 3,001–4,000 chars
The wizard shows your estimated credit cost in real time on Step 3 (Tools) and Step 5 (Review).
Plans
- Free: 3 credits on signup — never expire. No credit card required.
- Pro ($19/mo): 50 credits/month. Up to 25 unused credits roll over (max 75 total). Priority queue. Credit top-up packs available.
7. Understanding Your Security Score
Every FloMCP-generated server automatically receives a security score out of 100, calculated across 22 checks in 6 categories:
Input Validation
Schema strictness, type checking
Authentication
Env var usage, no hardcoded secrets
Error Handling
Graceful failures, no stack leaks
Rate Limiting
Abuse prevention patterns
SSRF Prevention
URL validation, metadata blocking
Dependency Safety
No known vulnerable packages
Scores 80–100 are production-ready. 60–79 need review. Below 60 should not be deployed to production without addressing the flagged issues.
9. FAQ
How many servers can I generate on the free plan?
Free accounts start with 3 credits — they never expire. Simple servers (≤3 tools, content ≤2,000 chars) cost 1 credit. Complex (4–10 tools or content 2,001–3,000 chars) cost 2 credits. Premium (11–20 tools or content 3,001–4,000 chars) cost 3 credits. Upgrade to Pro for 50 credits/month with 25-credit rollover.
What languages does FloMCP generate?
FloMCP currently generates TypeScript MCP servers using the official @modelcontextprotocol/sdk. Python support is on the roadmap.
Can I edit the generated code?
Yes — the generated code is yours. It's plain TypeScript with no proprietary dependencies. Edit, extend, and deploy it anywhere.
Does FloMCP store my generated servers?
Yes. Generated servers are saved in your dashboard (Dashboard → Servers). You can re-download the ZIP, view the full source code, check the security score, and access integration instructions at any time.
How do AI Tool Suggestions work?
On Step 3 (Tools), click Generate Tools. FloMCP sends your description to Claude and receives 3 suggested tool definitions — names, descriptions, and parameter schemas — pre-filled and ready to edit. You can regenerate as many times as you like.
Can I import an existing tool schema into the wizard?
Yes. On Step 3, click Import Schema and paste JSON. FloMCP supports four formats: OpenAI function-calling format, MCP inputSchema, FloMCP native, and plain arrays. Duplicates are detected and skipped automatically so you never get two tools with the same name.
What are Resources and Prompts?
Resources are static context blocks (docs, schemas, data files) attached to your server so Claude can reference them during tool calls. Prompts are reusable message templates. Both are optional and can be scoped to all tools or a specific tool. Drag & drop .txt, .md, or .json files in Step 4.
What if my API key is wrong in .env?
The generated server validates required env vars at startup and throws a clear error message specifying which variable is missing or empty.
Can I use FloMCP servers with Cursor or Windsurf?
Yes. Any tool that supports the MCP stdio transport works. Check your tool's documentation for the config file location and format.
How do I get support?
Use the Support section in your dashboard (Dashboard → Support). We typically respond within 24 hours.