Skip to main content

MCP

The Model Context Protocol (MCP) lets external AI assistants connect directly to Una and use its tools — queries, file system, actions, and more — from within their own interfaces. Once connected, an AI assistant can read data, run model queries, and interact with Una without leaving its native chat environment.

Before You Begin​

Authentication​

Depending on your client, you can authenticate with Una's MCP server using either OAuth or an API Token. OAuth is the recommended approach for interactive clients such as Claude Desktop and ChatGPT. API Tokens are better suited for developer tools like GitHub Copilot in VS Code.

Identifying Your Tenant​

Every request to the MCP server must identify which Una tenant to connect to. There are three ways to do this:

  • Request header — If your client supports custom headers, add a header named tenant with your tenant name as the value.
  • Scoped API Token — When generating an API Token, you can scope it to a specific tenant so no additional configuration is needed.
  • Default tenant — Set a default tenant in the Una Portal under Settings → API Tokens. Requests that do not include a tenant header will fall back to this default.

MCP URL​

Use the URL that corresponds to your Una environment:

PortalMCP Endpoint
portal.unasoft.apphttps://api.unasoft.app/mcp
portal.dev.unasoft.apphttps://api.dev.unasoft.app/mcp
portal.demo.unasoft.apphttps://api.demo.unasoft.app/mcp
portal.demoqa.unasoft.apphttps://api.demoqa.unasoft.app/mcp
portal.staging.unasoft.apphttps://api.staging.unasoft.app/mcp

You will need this URL when configuring the clients below.

Claude​

Claude Desktop MCP setup

Setup​

  1. Open Claude Desktop.
  2. Go to Settings → Connectors → Add Custom Connector.
  3. Fill in the connector details:
FieldValue
NameUna
URLYour MCP endpoint (see MCP URL)
Client ID3a9ea044-f3dd-48d5-a4e0-a092ff0568ec
Client Secretue28Q~TlOCwItWo8~2zWLNmspxrRzMx~Q42aIb8i

Once connected, Una's tools will appear in Claude's tool list. You can ask Claude to run Una queries, read files, or manage actions directly in the chat.

ChatGPT​

ChatGPT MCP app setup

Setup​

  1. Go to ChatGPT on the web and open Settings → Apps → Advanced Settings.
  2. Enable Development Mode.
  3. Go to Settings → Apps → Create App and fill in the app details:
FieldValue
NameUna
DescriptionConnect to Una for queries, file access, and actions
URLYour MCP endpoint (see MCP URL)
Client ID3a9ea044-f3dd-48d5-a4e0-a092ff0568ec
Client Secretue28Q~TlOCwItWo8~2zWLNmspxrRzMx~Q42aIb8i

GitHub Copilot​

GitHub Copilot connects to MCP servers via a configuration file in your VS Code workspace.

Setup​

  1. Generate an API Token in the Una Portal under Settings → API Tokens.

  2. In your project, open or create .vscode/mcp.json and add the Una server entry:

{
  "servers": {
    "una": {
      "type": "http",
      "url": "https://api.unasoft.app/mcp",
      "headers": {
        "x-api-token": "<your-api-token>",
        "tenant": "<your-tenant-name>"
      }
    }
  }
}

Replace <your-api-token> with the token you generated and <your-tenant-name> with your Una tenant name. Update the URL if you are connecting to a non-production environment.

  1. Reload VS Code.

Copilot Chat will now be able to call Una tools when answering questions that require data from your Una environment.