# Connect OpenAI API to HighLevel (GHL): Setup, Business Dashboard & App Integrations

> The OpenAI Responses API accepts HighLevel as a remote MCP tool with headers, and HighLevel lists OpenAI Playground as a supported client. Use it to build your own agent products on HighLevel data.

Last verified: October 25, 2026 · Source page: https://ghlfounders.com/agents/openai-api

## Key facts
| Fact | OpenAI API |
| --- | --- |
| Maker | OpenAI |
| Type | Model APIs |
| Launched | Remote MCP tools in the Responses API |
| Access & price | OpenAI API key |
| Where it lives | Your code; OpenAI Playground |
| Route to HighLevel | HighLevel MCP |
| Always on | If you schedule it |
| Approval controls | require_approval per tool call |
| Scheduling | Your own |
| Runs on | Your server |
| Official GHL MCP tools | Yes (direct) |

## How OpenAI API connects to HighLevel
- **HighLevel MCP:** Connects straight to HighLevel's official MCP server with a Private Integration Token kept on your machine or server.

## Set up OpenAI API with HighLevel in 3 steps
1. Get an OpenAI API key.
2. Add an mcp tool with HighLevel's URL and headers.
3. Run the request; the model discovers and calls HighLevel tools.

Python:

```
import os
from openai import OpenAI
client = OpenAI()
resp = client.responses.create(
  model="gpt-5",  # use your current model
  input="List my 5 newest HighLevel contacts",
  tools=[{
    "type": "mcp",
    "server_label": "highlevel",
    "server_url": "https://services.leadconnectorhq.com/mcp/",
    "headers": {
      "Authorization": "Bearer " + os.environ["GHL_PIT"],
      "locationId": os.environ["GHL_LOCATION_ID"]
    }
  }]
)
```

## Best HighLevel jobs for OpenAI API
- Custom agents inside your own SaaS
- Batch analysis of conversations
- Prototyping in Playground

## Use OpenAI API as your HighLevel business dashboard
Power your own dashboard product: call the model with HighLevel and other MCP tools attached, and render the results in your app or Founder OS.

## Connect HighLevel to your other business apps with OpenAI API
Business app reach: Any remote MCP server you attach

| App category | How OpenAI API reaches it |
| --- | --- |
| Remote MCP | Attach several MCP servers per request |
| Your data | Function calling for internal APIs |


## Permissions to grant
- contacts.readonly
- conversations.readonly

## Limits and watch-outs
- You handle hosting, scheduling and approvals.
- Keep tokens in server-side environment variables.

## FAQ
### Can I test HighLevel MCP in OpenAI Playground?
Yes. HighLevel lists OpenAI Playground as a supported client for remote MCP.

### Can OpenAI API connect HighLevel to Gmail, Slack, QuickBooks and my other apps?
Yes, within its reach: Any remote MCP server you attach. Remote MCP: Attach several MCP servers per request. Your data: Function calling for internal APIs.

### Can OpenAI API power a HighLevel business dashboard?
Power your own dashboard product: call the model with HighLevel and other MCP tools attached, and render the results in your app or Founder OS.

## Sources
- [How to Setup and Use the HighLevel MCP Server](https://help.gohighlevel.com/support/solutions/articles/155000005741)
- [HighLevel MCP developer docs](https://marketplace.gohighlevel.com/docs/other/mcp)