agents that can act on environmental data
Amentum's Model Context Protocol server lets agents query ocean, atmosphere, geomagnetic,
gravity, and aviation radiation data through tool calls. MCP is a standard that lets large
language models (LLMs) call tools (APIs) directly.
Use it when an agent needs to decide which environmental dataset to query, call the right
tool, and return a result grounded in deterministic backend data.
how MCP differs from APIs
- APIs: a developer writes the workflow, chooses endpoints, maps parameters, and
handles responses.
- MCP: the agent reads tool schemas, decides what to call and when, then uses the
same production APIs behind the scenes.
MCP does not replace APIs. It enables agent-driven use of the same scientific services your
software can already call directly.
example agent tasks
Marine routing
Input: "Plan a lower-risk route from Brisbane to Noumea for the next 48 hours."
Tools invoked: wave forecast, ocean currents, bathymetry.
Result: route notes with wave exposure, current assistance, and shallow-water warnings.
Aviation radiation exposure
Input: "Estimate radiation exposure for SYD to LAX at cruise altitude today."
Tools invoked: aviation radiation dose along flight path.
Result: dose estimate with flight context and the model inputs used.
architecture
User / Agent → LLM (Claude, GPT, Bedrock) → MCP server (Amentum) → Amentum
APIs
- Schema-driven: tools define required inputs, parameter types, and responses.
- No custom integration logic: standard MCP clients can discover and call tools.
- Deterministic backend data: outputs come from Amentum APIs, not generated guesses.
quick start
Works with standard MCP clients. Included with API subscription. To use the MCP server,
create an account in the developer portal, generate an API key, and subscribe to the
services your agent needs. The same API key is used for both the web APIs and MCP access.
Endpoint pattern:
https://mcp.amentum.io/mcp/{service}/
Available services:
ocean, atmosphere, avrad, geomag,
gravity.
Example client config:
{
"mcpServers": {
"amentum-ocean": {
"url": "https://mcp.amentum.io/mcp/ocean/",
"headers": {
"API-Key": "${AMENTUM_API_KEY}"
}
}
}
}
Example tool call:
User: "What is the ocean depth at 33.86 S, 151.21 E?"
Agent: calls app_api_bathymetry_endpoints_GEBCO_Bathymetry_point
Input: latitude=-33.86, longitude=151.21
Result: water depth from the Amentum bathymetry API, with structured response fields.