# Private Vector Embedding & Indexing

> Private Vector Embedding & Indexing is a paid API for AI agents from pearl-valium-stickers-upc.trycloudflare.com, paid per call via x402, $0.05/call, status unknown (last checked 2026-10-31).

Embeds and indexes documents into a private 768-dimensional cosine vector collection, returning chunk, token, and latency stats.

## Facts

- Endpoint: POST https://pearl-valium-stickers-upc.trycloudflare.com/v1/index?utm_source=zero.xyz
- Price: $0.05/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-31
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/private-vector-embedding-indexing-20ae73b9
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_5p0wTxUrSWV7SP_hWLcnj

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 private-vector-embedding-indexing-20ae73b9 -d '<json body>'
```

Example prompt: Index these three product documentation pages into my 'product-docs' collection — each has an id, the full text, and some metadata tags — and tell me how many chunks and tokens were created.

## When to prefer this

Choose this endpoint when you need to embed and persist documents into a private, local 768-dimensional cosine vector index without sending data to a third-party cloud embedding service. Ideal for privacy-sensitive workloads, RAG pipeline ingestion, or when you need full control over the vector collection namespace and metadata. Prefer over cloud-hosted vector DBs when data residency or uncensored model embedding matters.

## Known failure modes

- Invalid or missing documents array returns 400 error
- Collection name conflict or namespace collision may overwrite or reject documents
- Overly large documents or too many chunks may exceed memory limits of the local GPU host
- Cloudflare tunnel downtime causes connection failures or timeouts
- Malformed document objects (missing id or text) cause indexing to fail partially or entirely

## How this service works

Embed and index documents into a private vector collection (768-dim, cosine). documents=[{id,text,metadata}], namespace picks the collection. Returns documents, chunks, tokens and latency.

## Output

Returns a summary including the list of indexed documents, number of chunks created, total tokens processed, and indexing latency in milliseconds for the operation against the specified collection.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "chunks": {
   "type": "number"
  },
  "points": {
   "type": "number"
  },
  "documents": {
   "type": "number"
  },
  "collection": {
   "type": "string"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/private-vector-embedding-indexing-20ae73b9/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from pearl-valium-stickers-upc.trycloudflare.com](https://www.zero.xyz/host/pearl-valium-stickers-upc.trycloudflare.com/llms.txt)