MCP Server

Ocean.io MCP

Ocean.io MCP is a server that connects Ocean.io data directly to AI assistants. It allows large language model based tools to search, export and reason over Ocean.io company and people data in real time.

By connecting Ocean.io MCP to tools like Claude and Cursor, AI assistants can access Ocean.io data directly inside chat interfaces. This removes the need to build custom integrations for each tool and makes it easier to use Ocean data in everyday workflows.

Ocean.io MCP is designed for interactive use, exploration, bulk export and agent based automation.

Availability

Ocean.io MCP is available on plans under our new unified credit offering.

If you are unsure whether your current plan includes MCP access, please contact Ocean.io support to learn more about eligibility and upgrade options.

Prerequisites

To connect to the remote Ocean.io MCP server you'll need an active Ocean.io API key.

All requests made through the MCP server use the same API credits as regular Ocean.io API requests.

Integration with AI assistants

Ocean.io MCP works with any AI assistant that supports the Model Context Protocol.

When setting up the MCP server, replace the api key placeholder with your Ocean.io API key.

After completing the setup, make sure to refresh or reload the AI assistant so the new MCP server becomes available.

Claude Desktop

To connect Ocean.io MCP to Claude Desktop, follow these steps.

First, locate your Claude Desktop configuration file named claude_desktop_config.json.

Add the following configuration to the file and replace the api key placeholder with your Ocean.io API token.

{
  "mcpServers": {
    "ocean_data_api": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://api.ocean.io/mcp/?api-token=<api_key>"
      ]
    }
  }
}

Save the file and restart Claude Desktop.

After restarting, refresh the available tools. Ocean.io MCP should now be connected and ready to use.

Cursor

Cursor has built-in support for MCP servers.

1. Go to Cursor → Settings → Cursor Settings → MCP
2. Click Add new global MCP server
3. Paste the following configuration and replace the api key placeholder with your Ocean.io API token.

{
  "mcpServers": {
    "ocean_data_api": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://api.ocean.io/mcp/?api-token=<api_key>"
      ]
    }
  }
}

Save the configuration and restart Cursor if needed.

Once refreshed, Ocean.io MCP tools will be available inside Cursor.

Available tools

Ocean.io MCP exposes the following tools to AI assistants.

list_company_fields This tool returns all available company fields that can be requested or filtered on.

The output of this tool does not change. It should only be called once per session. Calling it more than once in the same session can reduce performance and harm the user experience.

list_people_fields This tool returns all available people fields that can be requested when searching for people.

The output is static and should only be called once per session.

list_industries This tool returns all available Ocean.io industries that can be used in company and people filters.

The output is static and should only be called once per session.

list_linkedin_industries This tool returns all available LinkedIn industry values that can be used for filtering.

The output does not change and should only be called once per session.

search_companies This tool is used to search for companies.

It supports filtering by company attributes and also supports lookalike search using an array of company domains to find similar companies.

Search results are paginated.

search_people This tool is used to search for people.

It supports filters such as job title, seniority, department, location, skills, and company information. It also supports lookalike search to find people similar to a given set of profiles or companies.

Search results are paginated.

export_companies This tool exports companies to a CSV file and returns a URL where the file can be downloaded.

It accepts up to 10,000 company domains per request.

Exports consume API credits at a rate of 0.2 credits per successfully exported company. Credits are only charged for companies that are successfully exported. For example: 5 companies exported = 1.0 credit 25 companies exported = 5.0 credits

AI assistants must confirm the estimated credit cost with the user before executing an export.

export_people This tool exports people to a CSV file using their LinkedIn handles or URLs and returns a download URL for the generated CSV file.

It accepts up to 10,000 LinkedIn IDs per request.

Exports consume API credits at a rate of 0.2 credits per successfully exported person. Credits are only charged for people that are successfully exported. For example: 5 people exported = 1.0 credit 25 people exported = 5.0 credits

AI assistants must confirm the estimated credit cost with the user before executing an export.

Tool behavior notes

Some tools return static reference data. These tools are idempotent and should only be called once per session. AI assistants are expected to cache the results and reuse them.

Search tools return paginated results. Each request returns a limited number of results and additional pages must be fetched separately.

Export tools generate a CSV file in a single operation and return a download URL. Unlike paginated search, exports do not require iterating through result pages. However, exports consume API credits based on the number of successfully exported records and require explicit user confirmation before execution.

Usage example

Below is an example of how Ocean.io MCP can be used in an AI assistant.

Example request

Find Chicago-headquartered companies similar to pandadoc.com and export them to CSV.

Example answer

Here are 5 companies similar to pandadoc.com HQed in Chicago.

1. OneSpan – Enterprise-grade e-signature and digital agreement security platform focused on regulated and high-risk transactions.
2. Mediafly – Sales enablement platform that helps revenue teams deliver interactive content and close deals faster.
3. Conga – Contract lifecycle management and document automation software powering complex revenue workflows.
4. Showpad – Sales enablement solution that equips sales teams with content, training, and buyer engagement tools.
5. ActiveCampaign – Sales and marketing automation platform for customer engagement, pipelines, and lifecycle messaging.

Export ready. This export will consume 1.0 credit (5 companies × 0.2 credits). Please confirm if you would like me to proceed with generating the CSV file.

Once confirmed, the assistant calls the export_companies tool and returns a download URL for the generated CSV file.