A real MCP session: the agent discovers the suite via suite-registry.json, selects a tool, constructs a typed invocation, and receives a deterministic JSON response. Zero server calls. Zero inference. Client-side replay.
Five discrete steps: no magic, no inference. Here is exactly what each phase does.
suite-registry.json at the root. 351 tools, typed schemas, MCP v1+ compliant manifest entries. No API key needed: the registry is public and machine-readable.manifest.json with a full inputSchema (JSON Schema Draft-07). The agent constructs a validated call: no prompt engineering, no natural-language parsing required.These tools appear in real agent invocation demos. Open any one and interact directly: client-side, zero PII.
Point your MCP-compatible agent at the AINumbers server endpoint. The registry loads once; tool schemas resolve on demand.
Connect once. Discover 351 tools via the registry. Each tool's manifest.json provides the full inputSchema, outputSchema, and policy_mandate envelope spec.
{
"registry_version": "1.0.0",
"mcp_version": "1+",
"tool_count": 351,
"categories": [
"fraud-risk", "payments", "aml", "compliance",
"esg", "credit", "capital", "liquidity",
"rbe-deterministic", "iso20022"
],
"tools": [
{
"id": "fraud-score-simulator",
"category": "fraud-risk",
"manifest": "https://ainumbers.co/tools/04-fraud-score-simulator/manifest.json",
"ap2": true,
"deterministic": true
},
"... 350 more tools"
]
}