# PennyRail Large Embedding Service

> PennyRail Large Embedding Service is a paid API for AI agents from pennyrail.vercel.app, paid per call via x402, $0.01/call, status unknown (last checked 2026-09-15).

Generates large-scale dense vector embeddings from an input object via a paid x402 micro-settlement endpoint

## Facts

- Endpoint: POST https://pennyrail.vercel.app/api/p/standard/ai.embed-large--embed-large
- Price: $0.01/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/pennyrail-large-embedding-service-97386e57
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_xo4Cy3vGbTzcvu8CGfgsh

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 pennyrail-large-embedding-service-97386e57 -d '<json body>'
```

Example prompt: Embed this product description using the large embedding model so I can store it in my vector database for semantic search: 'Lightweight hiking boots with waterproof membrane and Vibram outsole, ideal for multi-day trails.'

## When to prefer this

Choose this endpoint when you need large-model-quality embeddings with per-call micro-payment pricing via x402/USDC, especially in agentic pipelines where you want pay-as-you-go embedding without API key management. Prefer it over smaller embedding endpoints when semantic fidelity and vector dimensionality matter for downstream retrieval quality.

## Known failure modes

- Missing or malformed 'input' field returns a validation error
- Payment not settled via x402 protocol results in 402 Payment Required response
- Input object structure not understood by embedding model may return an empty or error response
- Network timeout on Vercel serverless cold start may cause latency spikes
- Overly large input may exceed model token limits

## How this service works

Machine-readable settlement service

## Output

Returns a JSON object containing the dense vector embedding produced by the large embedding model for the provided input, suitable for downstream semantic search, clustering, or similarity computation tasks.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object",
 "additionalProperties": true
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/pennyrail-large-embedding-service-97386e57/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from pennyrail.vercel.app](https://www.zero.xyz/host/pennyrail.vercel.app/llms.txt)