# Eval Engine API — Pay-per-call AI Evaluation

> Eval Engine API — Pay-per-call AI Evaluation is a paid API for AI agents from eval.zuluworksai.com, paid per call via x402, $0.005/call, status down (last checked 2026-09-15).

Scores LLM outputs, agent trajectories, and model responses against benchmark rubrics for automated AI quality evaluation

## Facts

- Endpoint: POST https://eval.zuluworksai.com/mcp
- Price: $0.005/call
- Payment: x402
- Status: down
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/eval-engine-api-pay-per-call-ai-evaluation-6eb8de23
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_PEdE-gVz5-5F0_d-usfKV

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 eval-engine-api-pay-per-call-ai-evaluation-6eb8de23 -d '<json body>'
```

Example prompt: Score this LLM response against my accuracy rubric: the model said 'Paris is the capital of Germany' when asked about European capitals — I want to know how it performs on factual correctness and give me a numeric quality score.

## When to prefer this

Choose this endpoint when you need pay-per-call, on-demand AI evaluation without committing to a subscription — ideal for CI/CD pipelines, spot-checking model outputs, or low-volume evaluation tasks. Best suited for teams that want to pay only for evals they run ($0.005 per call via USDC on Base) and need MCP-compatible tooling that integrates with agent frameworks. Prefer over batch eval platforms when you need real-time scoring in an agent loop.

## Known failure modes

- Payment failure if insufficient USDC balance on Base — returns 402 Payment Required
- Invalid JSON-RPC format causes parse error response with error object
- Missing required fields (id, method, jsonrpc) returns validation error
- Unknown MCP method returns method-not-found error
- Malformed agent trajectory or LLM output may return low-confidence evaluation or error
- Network timeout on complex trajectory evaluations

## How this service works

Pay-per-call AI evaluation engine. Score LLM outputs, agent trajectories, and model responses against benchmark rubrics. $0.005 per eval via x402 USDC on Base. Free trial available.

## Output

Returns a JSON-RPC response object containing evaluation results including scores, rubric assessments, and quality ratings for the submitted LLM output or agent trajectory. The result field contains the structured evaluation with numeric scores and qualitative feedback per benchmark dimension.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "id",
  "method",
  "jsonrpc"
 ],
 "properties": {
  "id": {
   "type": "string"
  },
  "method": {
   "type": "string",
   "description": "MCP method (tools/call, tools/list, etc.)"
  },
  "params": {
   "type": "object"
  },
  "jsonrpc": {
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "id": {
   "type": "string"
  },
  "error": {
   "type": "object"
  },
  "result": {
   "type": "object"
  },
  "jsonrpc": {
   "type": "string"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/eval-engine-api-pay-per-call-ai-evaluation-6eb8de23/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from eval.zuluworksai.com](https://www.zero.xyz/host/eval.zuluworksai.com/llms.txt)