INTEGRATIONS REAL-TIME DATA

Connect AI to Any System

Enable your AI agents to query databases and call APIs in real-time

MCP (Model Context Protocol) servers bridge your AI agents with external systems. From PostgreSQL databases to REST APIs, give your agents direct access to live data and the ability to take actions across your entire tech stack.

10+

Data sources

<100ms

Response time

99.9%

Uptime

Code required

Video Tutorial

Why Connect External Systems?

Real-Time Data

Access live information instead of stale knowledge. Query current inventory, customer status, and system metrics.

Take Actions

Create tickets, update records, send notifications. Turn conversations into actual business outcomes.

Enterprise Security

Secure connections with credential management, VPN support, and audit logging for compliance.

No Code Required

Visual configuration for database connections and API endpoints. No programming knowledge needed.

MCP Server Types

Database Server

API Server

FLEXIBLE

Integrate with REST APIs for external service interactions.

REST API endpoints
Authentication headers
Request/response mapping
Rate limiting
Error handling

Create a Database Server

Step 1: Configure Connection

Navigate to MCP ServersGo to the MCP Servers section in your dashboard
Click "New Database Server"Choose database type from supported options
Enter connection detailsHost, port, database name, and credentials
Test connectionVerify connectivity before saving

Step 2: Configure Permissions

Select tablesChoose which tables agents can access
Set operation typesRead-only, read-write, or custom permissions
Define query limitsMaximum rows returned and timeout settings
Review schemaConfirm table structure and relationships

Step 3: Test & Deploy

Run test queriesValidate configuration with sample queries
Check performanceMonitor response times and resource usage
Deploy serverActivate the MCP server for agent use
Monitor healthSet up alerts for connection issues

Step 4: Assign to Agents

Select agentsChoose which agents can use this server
Configure accessSet permissions per agent if needed
Test integrationVerify agents can query successfully
Document usageCreate examples for prompt engineering

Example Configuration

{
  "name": "Customer Database",
  "type": "postgresql",
  "connection": {
    "host": "db.company.com",
    "port": 5432,
    "database": "customers",
    "username": "ai_agent",
    "password": "secure_password"
  },
  "permissions": {
    "tables": ["customers", "orders", "products"],
    "operations": ["SELECT", "INSERT", "UPDATE"],
    "max_rows": 1000,
    "timeout": 30
  },
  "security": {
    "ssl_mode": "require",
    "connection_limit": 10
  }
}

Security Warning

Always use dedicated database users with minimal required permissions. Never use admin credentials for MCP servers. Enable SSL/TLS for all database connections.

Create an API Server

Choose API typeREST API with authentication method (API key, OAuth, Basic auth)
Configure base URLSet the root endpoint for all API calls
Define endpointsMap specific API routes your agents will use
Set up authenticationConfigure headers, tokens, or credential management
Test endpointsValidate all configured API calls work correctly
Deploy and monitorActivate the server and set up health monitoring

Freshdesk Integration Example

{
  "name": "Freshdesk Support API",
  "type": "rest_api",
  "base_url": "https://company.freshdesk.com/api/v2",
  "authentication": {
    "type": "api_key",
    "header": "Authorization",
    "value": "Basic {{encoded_api_key}}"
  },
  "endpoints": [
    {
      "name": "get_tickets",
      "method": "GET",
      "path": "/tickets",
      "description": "Retrieve support tickets"
    },
    {
      "name": "create_ticket",
      "method": "POST", 
      "path": "/tickets",
      "description": "Create a new support ticket"
    },
    {
      "name": "update_ticket",
      "method": "PUT",
      "path": "/tickets/{id}",
      "description": "Update ticket status or details"
    }
  ]
}

Popular Integrations

Freshdesk

Support tickets

Jira

Issue tracking

Salesforce

CRM data

PostgreSQL

Database queries

HubSpot

Marketing & sales

Zendesk

Customer support

Using MCP Servers in Agents

Create or edit an agentGo to the AI Agents section and select your agent
Add MCP server toolsSelect which MCP servers this agent can use
Configure permissionsSet read-only, read-write, or custom access levels
Update agent instructionsAdd context about available data sources and operations
Test with sample queriesVerify the agent can successfully access and use the data

Pro Tip

Agents automatically discover available MCP tools and their capabilities. Include examples in your agent instructions to guide how tools should be used for best results.

Best Practices

Security

Use dedicated service accounts
Implement least-privilege access
Enable SSL/TLS encryption
Regular credential rotation
Monitor access logs
IP whitelist when possible

Performance

Set appropriate query limits
Use connection pooling
Index frequently queried fields
Monitor response times
Cache static data when possible
Optimize SQL queries

Reliability

Set up health monitoring
Configure timeout values
Implement retry logic
Plan for failover scenarios
Test backup connections
Document troubleshooting steps

Troubleshooting

Connection timeout errors

Cause: Network latency or database overload

Solution: Increase timeout values, check network connectivity, optimize queries, or scale database resources

Authentication failures

Cause: Incorrect credentials or expired tokens

Solution: Verify credentials, check token expiration, ensure proper permission levels, and test manually

Query performance issues

Cause: Unoptimized queries or missing indexes

Solution: Add database indexes, reduce query complexity, implement row limits, and use query analysis tools

Agent can't access MCP tools

Cause: Missing permissions or server deployment issues

Solution: Check agent MCP server assignments, verify server status, review permissions, and restart if needed

Ready to Automate with Workflows?

Combine MCP servers with workflows to create powerful automated processes.

Learn About Workflows Contact Sales