MCP for AI Agents

The Oviond API exposes a Model Context Protocol (MCP) endpoint, allowing AI assistants like Claude, Cursor, and other MCP-compatible clients to discover and query integration data through structured tools.

Endpoint

https://api.oviond.com/mcp

The endpoint uses the Streamable HTTP transport and speaks JSON-RPC 2.0.

Available Tools

`list_integrations`

List all available integrations with their IDs, names, categories, auth types, and data views.

Parameters: None

`get_integration`

Get detailed information about a specific integration including all metrics and dimensions.

Parameter	Type	Required	Description
`id`	string	Yes	The integration ID (e.g. `"google-analytics"`, `"stripe"`)

`query_integration`

Query data from an integration with specified metrics, dimension, and date range.

Parameter	Type	Required	Description
`id`	string	Yes	The integration ID
`credentials`	object	Yes	Auth credentials object
`profile`	object	Yes	Profile/account selector
`dataView`	string	Yes	Data view to query
`metrics`	string[]	Yes	Metric IDs to fetch
`dimension`	string	No	Single dimension ID
`dimensions`	string[]	No	Multiple dimension IDs
`dateRange`	object	Yes	`{ current: { startDate, endDate }, previous: { startDate, endDate } }` (YYYY-MM-DD)
`timezone`	string	No	IANA timezone (defaults to `"UTC"`)

`test_connection`

Test if credentials are valid for an integration.

Parameter	Type	Required	Description
`id`	string	Yes	The integration ID
`credentials`	object	Yes	Auth credentials object
`profile`	object	No	Profile/account selector

Connecting from MCP Clients

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "oviond": {
      "type": "streamable-http",
      "url": "https://api.oviond.com/mcp"
    }
  }
}

Cursor

Add to your .cursor/mcp.json:

{
  "mcpServers": {
    "oviond": {
      "type": "streamable-http",
      "url": "https://api.oviond.com/mcp"
    }
  }
}

Claude Code

claude mcp add oviond --transport http https://api.oviond.com/mcp

Testing with curl

You can verify the endpoint with a JSON-RPC initialize handshake:

curl -X POST https://api.oviond.com/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "initialize",
    "params": {
      "protocolVersion": "2025-03-26",
      "capabilities": {},
      "clientInfo": { "name": "curl", "version": "1.0.0" }
    }
  }'

A successful response returns the server info and available capabilities:

{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "protocolVersion": "2025-03-26",
    "capabilities": { "tools": {} },
    "serverInfo": { "name": "Oviond Integration API", "version": "1.0.0" }
  }
}

MCP for AI Agents

On this page