AI assistants (MCP)

The Voydar MCP server lets AI assistants (Claude, Cursor, VS Code, ChatGPT through the OpenAI API, and any other MCP client) look up vessels, tracks, ports, companies and sanctions for you. It is hosted: there is nothing to install.

Server address (Streamable HTTP):

https://www.seadar.io/mcp

Sign in with OAuth: assistants that support it (Claude.ai, ChatGPT, the MCP inspector) send you to Voydar to approve the connection, and it then uses your API credits. Or connect with your Voydar API key in the Authorization header, exactly as for the REST API; the free Sandbox key works. Connecting and listing tools are free. Each tool call is one API request with the same credits, rate limit and rights: withheld fields stay withheld, and every result carries the attribution to show and your plan's commercial-use scope. Connected apps are listed, and can be disconnected, under Account, API keys.

Tools (all read-only):

search_entities
get_vessel
get_vessel_track
get_port
get_port_calls
get_company
list_sanctioned_vessels
screen_vessels
get_chokepoint_transits
get_usage
get_my_feed

It also offers a resource explaining data rights and attribution (seadar://docs/rights) and two prompts: a vessel due-diligence brief and a port activity brief.

Add Voydar to your assistant

Claude.ai and the Claude apps: Settings, Connectors, Add custom connector, and enter the server address. Claude sends you to Voydar to approve.
ChatGPT: in Settings, Apps and connectors, turn on developer mode, create a connector with the server address and OAuth authentication, and approve on Voydar.

Claude Code (terminal):

claude mcp add --transport http seadar https://www.seadar.io/mcp \
  --header "Authorization: Bearer $SEADAR_API_KEY"

Claude Desktop (claude_desktop_config.json, through the mcp-remote bridge):

{
  "mcpServers": {
    "seadar": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://www.seadar.io/mcp", "--header", "Authorization:${SEADAR_AUTH}"],
      "env": { "SEADAR_AUTH": "Bearer sk_sdr_…" }
    }
  }
}

Cursor (~/.cursor/mcp.json):

{
  "mcpServers": {
    "seadar": {
      "url": "https://www.seadar.io/mcp",
      "headers": { "Authorization": "Bearer sk_sdr_…" }
    }
  }
}

VS Code (.vscode/mcp.json; asks for the key once):

{
  "inputs": [{ "type": "promptString", "id": "seadar-key", "description": "Voydar API key", "password": true }],
  "servers": {
    "seadar": {
      "type": "http",
      "url": "https://www.seadar.io/mcp",
      "headers": { "Authorization": "Bearer ${input:seadar-key}" }
    }
  }
}

OpenAI API (Responses API): add Voydar as a remote MCP tool with your key:

curl https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $OPENAI_API_KEY" -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5",
    "tools": [{
      "type": "mcp",
      "server_label": "seadar",
      "server_url": "https://www.seadar.io/mcp",
      "headers": { "Authorization": "Bearer '"$SEADAR_API_KEY"'" },
      "require_approval": "never"
    }],
    "input": "Is the vessel IMO 9811000 on any sanctions list?"
  }'

Replace sk_sdr_… with your key, or keep it in an environment variable. A key in a config file is a secret: do not commit it or share the file.