# BGE-M3 Text Embeddings API

> BGE-M3 Text Embeddings API is a paid API for AI agents from onchaincheck.dev, paid per call via x402, $0.002/call, status unknown (last checked 2026-11-12).

Generates multilingual BGE-M3 text embeddings (1,024 dimensions) for one or up to 32 texts per call, supporting semantic search, clustering, and similarity tasks.

## Facts

- Endpoint: POST https://onchaincheck.dev/ai/embeddings?utm_source=zero.xyz
- Price: $0.002/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-12
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/bge-m3-text-embeddings-api-75f5abff
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_5S9-ZU9GAGXnw2Kiq0RWB

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 bge-m3-text-embeddings-api-75f5abff -d '<json body>'
```

Example prompt: Embed these three product descriptions into vectors using BGE-M3 so I can compare their semantic similarity: 'wireless noise-cancelling headphones', 'bluetooth over-ear audio headset', 'wired studio monitor earphones'.

## When to prefer this

Choose this endpoint when you need high-quality multilingual embeddings (supporting 100+ languages) with a 1,024-dimensional output from the BGE-M3 model, especially when batching up to 32 texts in a single call for efficiency. Prefer this over OpenAI or Cohere embeddings when you want on-chain micropayment billing via x402 or when multilingual coverage and BGE-M3's retrieval performance are priorities.

## Known failure modes

- Input exceeds 32,000 total characters — returns a 400 error
- Array contains more than 32 texts — returns a 400 error
- Empty text string submitted — returns a 400 or validation error
- Payment not provided or insufficient USDC — returns a 402 Payment Required
- Service temporarily unavailable — returns a 503 error

## How this service works

Text embeddings with bge-m3 (1,024 dimensions, multilingual): one text, or up to 32 texts in one call (POST texts array), up to 32,000 characters in all. For semantic search, clustering and similarity.

## Output

Returns 1,024-dimensional float vectors (BGE-M3 embeddings) for each input text. If a single text was submitted, returns one embedding vector; if an array was submitted, returns an array of embedding vectors in the same order as the inputs. Vectors can be used directly for cosine similarity, dot-product comparison, nearest-neighbor search, or as input to clustering algorithms.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "text": {
   "type": "string",
   "pattern": "^[\\s\\S]+$",
   "maxLength": 32000,
   "description": "Text to embed (or texts: an array of up to 32 strings)"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/bge-m3-text-embeddings-api-75f5abff/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from onchaincheck.dev](https://www.zero.xyz/host/onchaincheck.dev/llms.txt)
