# ARBITER Semantic Ranking

> ARBITER Semantic Ranking is a paid API for AI agents from api.grip.fyi, paid per call via x402, $0.01/call, status unknown (last checked 2026-09-14).

Ranks a fixed list of candidate strings against a query representing current state, intent, or perspective using deterministic semantic scoring.

## Facts

- Endpoint: POST https://api.grip.fyi/v1/compare
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/arbiter-semantic-ranking-ae946b98
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_ylI5ncsinX3O9bp-FvLfG

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 arbiter-semantic-ranking-ae946b98 -d '<json body>'
```

Example prompt: Rank these 5 response strategies — ['apologize and refund', 'escalate to manager', 'offer store credit', 'request more info', 'close the ticket'] — against the current intent 'angry customer reporting a missing order', and return the top 3 most appropriate options.

## When to prefer this

Choose this endpoint when you have a fixed, enumerable set of candidate strings and need to deterministically rank them against a query, intent, or current state — especially in agentic control-flow scenarios where you need reproducible, explainable ordering rather than generative output. Prefer this over embedding-based vector search when the candidate set is small and known in advance, and over LLM-based ranking when determinism and cost predictability matter.

## Known failure modes

- Empty candidates array returns a validation error
- top_k larger than candidates array length may return fewer results
- Highly ambiguous or very short query strings may produce low-confidence rankings
- Non-string entries in candidates array may cause schema validation failures
- Network timeouts or payment failures (x402) may prevent results from being returned

## How this service works

ARBITER deterministic semantic control-flow and arbitration: measure and rank a fixed candidate field against current state, intent, or perspective.

## Output

Returns a ranked list of up to top_k candidate strings ordered by their deterministic semantic relevance to the query, along with their associated scores or rankings indicating closeness of fit to the stated intent or state.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "query": {
   "type": "string",
   "description": "Current state, intent, question, or perspective."
  },
  "top_k": {
   "type": "integer",
   "default": 3,
   "minimum": 1
  },
  "use_freq": {
   "type": "boolean",
   "default": true
  },
  "candidates": {
   "type": "array",
   "items": {
    "type": "string"
   },
   "minItems": 1,
   "description": "Fixed candidate field to measure and rank."
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/arbiter-semantic-ranking-ae946b98/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.grip.fyi](https://www.zero.xyz/host/api.grip.fyi/llms.txt)