OpenAI Agent Builder

Overview

OpenAI Agent Builder allows you to create custom AI agents with access to external tools and data sources. By connecting Lemonado via MCP, your agents can query databases, SaaS tools, and analytics platforms in real-time.

Prerequisites

• OpenAI account with access to Agent Builder
• Active Lemonado AI Connector with MCP endpoint configured
• API access token (generated from Lemonado)

Setup Instructions

Step 1: Generate Lemonado Access Token

1. Log in to your Lemonado dashboard
2. Navigate to your AI Connector settings
3. Click Generate Access Token
4. Copy the bearer token (it will only be shown once)
5. Save it securely - this token never expires unless regenerated

Step 2: Configure OpenAI Agent Builder

1. Go to your agent in OpenAI Agent Builder
2. Click Add Tool → MCP Server
3. Choose + Server in the top right
4. Fill in the server details:

5. Paste your Lemonado bearer token in the authentication field
6. Click Connect to save the configuration

Step 3: Test the Connection

1. In your agent's chat interface, try a test query:

"List all available data sources from Lemonado"

2. If successful, your agent will return information from your Lemonado AI Connector

Capabilities

With Lemonado connected to your OpenAI agent, you can:

Data Access

• Query any connected database or SaaS platform
• Access real-time synchronized data
• Filter and search across multiple data sources
• Retrieve specific records and datasets

Analytics & Reporting

• Generate custom reports from your data
• Perform aggregations and calculations
• Analyze trends and patterns
• Create data visualizations

Multi-Source Queries

• Combine data from multiple sources
• Cross-reference information
• Build comprehensive dashboards
• Track metrics across platforms

Security & Best Practices

Token Management

• Never share your access token publicly or commit it to version control
• Regenerate tokens immediately if compromised
• Use separate tokens for different agents or environments
• Monitor usage in your Lemonado audit logs
• Revoke access by regenerating the token

Access Control

• Configure data permissions in your Lemonado AI Connector settings
• Use data filtering to limit what data your agents can access
• Set up read-only access for security
• Monitor all agent queries in real-time

Rate Limiting

• OpenAI enforces rate limits on API calls
• Plan your agent's query patterns accordingly
• Consider caching frequently accessed data
• Monitor your Lemonado usage dashboard

Advanced Configuration

Multiple AI Connectors

You can add multiple Lemonado AI Connectors to a single agent:

1. Create separate AI Connectors in Lemonado for different data sources
2. Generate a unique access token for each connector
3. Add each as a separate MCP server in OpenAI Agent Builder
4. Give each a descriptive name (e.g., "Lemonado-Sales", "Lemonado-Analytics")

Custom Agent Instructions

Optimize your agent's data access by including specific instructions:

When querying Lemonado:
1. Always specify date ranges for time-series data
2. Use filters to limit result sets
3. Request only the fields you need
4. Handle empty results gracefully

Troubleshooting

Authentication Failed

1. Verify token is correct

   • Check for extra spaces or characters
   • Ensure you copied the complete token
   • Regenerate if unsure

2. Token expired or revoked

   • Generate a new token in Lemonado
   • Update the token in OpenAI Agent Builder
   • Note: Tokens don't expire unless manually regenerated

Connection Issues

• Check MCP server URL: Ensure https://mcp.lemonado.io/mcp is entered correctly
• Verify AI Connector status: Confirm your AI Connector is active in Lemonado
• Test from Lemonado: Try querying your data directly in Lemonado first

No Data Returned

1. Verify data sources are connected in your Lemonado AI Connector
2. Check data sync status - ensure recent sync was successful
3. Review permissions - confirm the AI Connector has access to required data
4. Test query syntax - verify your agent is using correct query format

Next Steps

• Explore AI Connector configuration options
• Learn about data source connections
• Review security best practices
• Join our community for agent templates and examples

Need Help?

• Visit our documentation
• Contact support at support@lemonado.io
• Check the MCP documentation