# Testing Guide
## Overview
This guide provides comprehensive testing strategies for the MCP server, from local development to production validation.
## Testing Strategy
### 1. Unit Testing (Manual)
Test individual components in isolation
### 2. Integration Testing
Test MCP protocol compliance and OpenAI API integration
### 3. End-to-End Testing
Test with real MCP clients (Claude Desktop, Roo)
### 4. Performance Testing
Validate response times and resource usage
## Local Development Testing
### 1. Start Development Server
```bash
# Start Wrangler dev server
npm run dev
# Server should start on http://localhost:8787
```
### 2. Health Check Test
```bash
# Test basic connectivity
curl http://localhost:8787/
# Expected response:
{
"status": "OpenAI Vector Store MCP Server Running",
"version": "2.0.0",
"endpoint": "/mcp/{api-key}"
}
```
### 3. Authentication Tests
#### Valid API Key Test
```bash
curl -X POST http://localhost:8787/mcp/your-test-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {
"protocolVersion": "2024-11-05",
"capabilities": {}
},
"id": 1
}'
# Expected: 200 OK with initialize response
```
#### Invalid API Key Test
```bash
curl -X POST http://localhost:8787/mcp/invalid-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {},
"id": 1
}'
# Expected: 401 Unauthorized
{
"jsonrpc": "2.0",
"error": {
"code": -32600,
"message": "Unauthorized: Invalid API key"
},
"id": null
}
```
## MCP Protocol Testing
### 1. Initialize Method Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {
"protocolVersion": "2024-11-05",
"capabilities": {
"roots": {"listChanged": true},
"sampling": {}
}
},
"id": 1
}'
# Expected response:
{
"jsonrpc": "2.0",
"result": {
"protocolVersion": "2024-11-05",
"capabilities": {
"tools": {},
"logging": {}
},
"serverInfo": {
"name": "openai-vector-store-mcp",
"version": "2.0.0"
}
},
"id": 1
}
```
### 2. Tools List Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/list",
"id": 2
}'
# Expected: List of 4 vector store tools
```
### 3. Tools Call Tests
#### Vector Store List Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-list",
"arguments": {
"limit": 5,
"order": "desc"
}
},
"id": 3
}'
# Expected: JSON array of vector stores
```
#### Vector Store Create Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-create",
"arguments": {
"name": "Test Vector Store",
"expires_after_days": 7,
"metadata": {
"purpose": "testing",
"created_by": "mcp_test"
}
}
},
"id": 4
}'
# Expected: New vector store object with ID
```
#### Vector Store Get Test
```bash
# Use ID from create test above
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-get",
"arguments": {
"vector_store_id": "vs_abc123"
}
},
"id": 5
}'
# Expected: Vector store details
```
#### Vector Store Delete Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-delete",
"arguments": {
"vector_store_id": "vs_abc123"
}
},
"id": 6
}'
# Expected: Deletion confirmation
```
## Error Handling Tests
### 1. Invalid Method Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "invalid/method",
"id": 7
}'
# Expected: Method not found error (-32601)
```
### 2. Invalid Tool Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "invalid-tool",
"arguments": {}
},
"id": 8
}'
# Expected: Tool not found error
```
### 3. Missing Parameters Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-create",
"arguments": {}
},
"id": 9
}'
# Expected: OpenAI validation error
```
### 4. Invalid JSON Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{"invalid": json}'
# Expected: JSON parse error
```
## CORS Testing
### 1. Preflight Request Test
```bash
curl -X OPTIONS http://localhost:8787/mcp/your-api-key \
-H "Origin: https://example.com" \
-H "Access-Control-Request-Method: POST" \
-H "Access-Control-Request-Headers: Content-Type"
# Expected: 204 No Content with CORS headers
```
### 2. Cross-Origin Request Test
```bash
curl -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-H "Origin: https://example.com" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {},
"id": 1
}'
# Expected: Response with Access-Control-Allow-Origin: *
```
## Performance Testing
### 1. Response Time Test
```bash
# Test multiple requests and measure time
for i in {1..10}; do
time curl -s -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/list",
"id": '$i'
}' > /dev/null
done
# Expected: < 100ms per request (excluding OpenAI API time)
```
### 2. Concurrent Request Test
```bash
# Test 10 concurrent requests
for i in {1..10}; do
curl -s -X POST http://localhost:8787/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/list",
"id": '$i'
}' &
done
wait
# Expected: All requests succeed
```
## MCP Client Testing
### 1. Create Test Configuration
#### For Claude Desktop
Create `claude_desktop_config.json`:
```json
{
"mcpServers": {
"openai-vector-store-test": {
"command": "npx",
"args": [
"@modelcontextprotocol/server-fetch",
"http://localhost:8787/mcp/your-api-key"
]
}
}
}
```
#### For Roo
Create MCP settings:
```json
{
"mcpServers": {
"openai-vector-store-test": {
"command": "npx",
"args": [
"@modelcontextprotocol/server-fetch",
"http://localhost:8787/mcp/your-api-key"
]
}
}
}
```
### 2. Test MCP Client Integration
1. **Start your MCP server** (local or deployed)
2. **Configure MCP client** with test configuration
3. **Restart MCP client** to load new configuration
4. **Test tool availability** - tools should appear in client
5. **Test tool execution** - try creating/listing vector stores
### 3. Validate Tool Responses
Test each tool through the MCP client:
- `vector-store-list` - Should return formatted list
- `vector-store-create` - Should create and return new store
- `vector-store-get` - Should return store details
- `vector-store-delete` - Should confirm deletion
## Production Testing
### 1. Deploy to Cloudflare
```bash
# Deploy to production
npm run deploy
# Set production secrets
echo "your-openai-api-key" | wrangler secret put OPENAI_API_KEY
echo "your-mcp-api-key" | wrangler secret put MCP_API_KEY
```
### 2. Test Production Endpoint
```bash
# Replace with your actual worker URL
curl https://your-worker.workers.dev/
# Test MCP endpoint
curl -X POST https://your-worker.workers.dev/mcp/your-api-key \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {},
"id": 1
}'
```
### 3. Update MCP Client Configuration
```json
{
"mcpServers": {
"openai-vector-store": {
"command": "npx",
"args": [
"@modelcontextprotocol/server-fetch",
"https://your-worker.workers.dev/mcp/your-api-key"
]
}
}
}
```
## Automated Testing Script
Create `test-script.sh`:
```bash
#!/bin/bash
# Configuration
BASE_URL="http://localhost:8787"
API_KEY="your-test-api-key"
MCP_ENDPOINT="$BASE_URL/mcp/$API_KEY"
echo "Testing MCP Server..."
# Test 1: Health check
echo "1. Testing health endpoint..."
curl -s "$BASE_URL/" | jq .
# Test 2: Initialize
echo "2. Testing initialize..."
curl -s -X POST "$MCP_ENDPOINT" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "initialize",
"params": {"protocolVersion": "2024-11-05"},
"id": 1
}' | jq .
# Test 3: Tools list
echo "3. Testing tools/list..."
curl -s -X POST "$MCP_ENDPOINT" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/list",
"id": 2
}' | jq .
# Test 4: Vector store list
echo "4. Testing vector-store-list..."
curl -s -X POST "$MCP_ENDPOINT" \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "vector-store-list",
"arguments": {"limit": 5}
},
"id": 3
}' | jq .
echo "Testing complete!"
```
Make it executable and run:
```bash
chmod +x test-script.sh
./test-script.sh
```
## Troubleshooting Common Issues
### 1. Connection Refused
- Check if Wrangler dev server is running
- Verify port 8787 is not blocked
- Try `wrangler dev --port 8788` for different port
### 2. Authentication Errors
- Verify API key in URL matches MCP_API_KEY secret
- Check OpenAI API key is valid and has vector store permissions
- Ensure secrets are set correctly: `wrangler secret list`
### 3. CORS Errors
- Check browser developer tools for specific CORS issues
- Verify CORS headers are included in all responses
- Test with curl first to isolate CORS from other issues
### 4. OpenAI API Errors
- Test OpenAI API key directly with curl
- Check rate limits and quotas
- Verify API key has vector store permissions
### 5. MCP Client Issues
- Check MCP client logs for connection errors
- Verify client configuration syntax
- Test with simple HTTP client first
- Restart MCP client after configuration changes
This comprehensive testing approach ensures your MCP server works correctly across all scenarios and integrates properly with MCP clients.
