# MCP Usage Coverage

> MCP Usage Coverage is a paid API for AI agents from api.zfinia.com, paid per call via x402, $0.12/call, status unknown (last checked 2026-10-12).

Audits a transcript of MCP tool call recordings against a declared toolset manifest to identify unused tools, unused or constant arguments, and undeclared tool invocations.

## Facts

- Endpoint: POST https://api.zfinia.com/x402/v1/mcp-tool-input-coverage-audit?utm_source=zero.xyz
- Price: $0.12/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-12
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/mcp-usage-coverage-71c98409
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_HhhUrWfrGAGtme8rdFPvg

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability mcp-usage-coverage-71c98409 -d '<json body>'
```

Example prompt: Can you audit this batch of MCP tool call transcripts against our declared toolset manifest and tell me which declared tools were never called, which arguments were never sent, which arguments always carry the same value, and whether any calls hit tools we didn't declare?

## When to prefer this

Choose this endpoint when you need a passive, read-only audit of agent tool usage against a declared MCP manifest — especially useful after agent deployments to identify dead tool declarations, pruning candidates, hardcoded argument antipatterns, or manifest drift. Prefer this over manual log inspection or generic diff tools because it understands the MCP tool-call structure specifically and outputs a structured gap report rather than raw diffs.

## Known failure modes

- Empty calls array — minimum 1 item required, returns validation error
- Malformed tool call objects in the calls array — may produce incomplete or skewed coverage results
- Toolset manifest missing or null — undeclared-tool detection may be skipped or error
- Transcript contains over 1000 calls — exceeds maxItems limit, request rejected
- Mismatch between transcript schema and toolset format — may fail to correlate tools and arguments correctly

## How this service works

Find out which tools and arguments your agent actually uses. From a transcript of recorded calls, reports which declared tools were never called, which declared arguments were never sent, which arguments always carry the same value, and any call to a tool the manifest does not declare. Nothing is executed.

## Output

A structured coverage report listing: tools declared in the manifest that were never called, declared arguments that were never sent in any call, arguments that always carried an identical value across all calls (suggesting hardcoding), and any tool calls found in the transcript that reference tools absent from the manifest. No code is executed; the report is purely analytical.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "calls": {
   "type": "array",
   "items": {
    "type": "object"
   },
   "maxItems": 1000,
   "minItems": 1
  },
  "toolset": {}
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/mcp-usage-coverage-71c98409/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.zfinia.com](https://www.zero.xyz/host/api.zfinia.com/llms.txt)