# Vektorwerk Similarity Scorer

> Vektorwerk Similarity Scorer is a paid API for AI agents from vektor.halowerk.com, paid per call via x402, $0.001/call, status unknown (last checked 2026-09-13).

Embeds a query and up to 64 candidate strings locally, then returns cosine similarity scores and stable descending ranks for each candidate.

## Facts

- Endpoint: POST https://vektor.halowerk.com/similarity
- Price: $0.001/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-13
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/vektorwerk-similarity-scorer-e6af4354
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_7-T5fye5M9v-BXVgBEZyp

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 vektorwerk-similarity-scorer-e6af4354 -d '<json body>'
```

Example prompt: Score and rank these 5 candidate answers against my query 'how do I reset my password?' using the nomic model: 'click forgot password on the login page', 'update your profile settings', 'contact support by email', 'use the reset link sent to your inbox', 'log out and log back in'.

## When to prefer this

Prefer this endpoint when you need fast, local semantic similarity scoring or reranking of up to 64 candidates against a single query without relying on a third-party cloud embedding provider. Ideal for RAG pipelines needing a cheap ($0.001/call) reranking step, FAQ matching, or any scenario where cosine similarity over a small fixed candidate set is sufficient and data privacy or latency matters.

## Known failure modes

- More than 64 candidates provided — API rejects with validation error
- Empty candidates array (below minItems of 1) — validation error returned
- Invalid model name outside 'nomic' or 'bge_m3' enum — request rejected
- Query string is empty or missing — request fails validation
- Service unavailable or local Ollama runtime down — connection error or 5xx response

## How this service works

Embeds one query and up to 64 candidates locally, then returns cosine scores and stable descending ranks.

## Output

Returns a list of candidate strings paired with their cosine similarity scores relative to the query, sorted in stable descending order by score, so the most semantically relevant candidate appears first.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "model": {
   "enum": [
    "nomic",
    "bge_m3"
   ],
   "type": "string",
   "default": "nomic"
  },
  "query": {
   "type": "string"
  },
  "candidates": {
   "type": "array",
   "items": {
    "type": "string"
   },
   "maxItems": 64,
   "minItems": 1
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/vektorwerk-similarity-scorer-e6af4354/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from vektor.halowerk.com](https://www.zero.xyz/host/vektor.halowerk.com/llms.txt)