# NoJunk LLM Embeddings (BGE-M3)

> NoJunk LLM Embeddings (BGE-M3) is a paid API for AI agents from nojunk-llm.mick8511.workers.dev, paid per call via x402, $0.001/call, status unknown (last checked 2026-10-19).

Generates 1024-dimensional vector embeddings for up to 32 text strings per call using BGE-M3, paid per-request in USDC with no API key required.

## Facts

- Endpoint: POST https://nojunk-llm.mick8511.workers.dev/v1/embeddings?utm_source=zero.xyz
- Price: $0.001/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-19
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/nojunk-llm-embeddings-bge-m3-5b622e75
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_RkKthXDQZ2EC6W1P5Wg3A

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 nojunk-llm-embeddings-bge-m3-5b622e75 -d '<json body>'
```

Example prompt: Generate BGE-M3 embeddings for these two sentences: 'The quick brown fox jumps over the lazy dog' and 'A fast auburn fox leaps above a sleepy hound' — I need the 1024-dim vectors to compare their semantic similarity.

## When to prefer this

Choose this endpoint when you need BGE-M3 embeddings (multilingual, strong cross-lingual retrieval) with zero onboarding friction — no API key, no account, no rate limits — and you want to pay only when a valid result is returned, billed per-call in USDC on Base via x402. Ideal for agents, prototypes, or low-volume workloads where provisioning a traditional embedding API account is overkill. Also suitable when your stack already uses an OpenAI-compatible client and you just need to swap the base URL.

## Known failure modes

- Input array exceeds 32 strings — request rejected
- Individual string exceeds 8000 characters — may be truncated or rejected
- Payment via x402 fails or USDC balance insufficient — no charge, no result returned
- Invalid JSON body or missing 'input' field — request fails
- Model field set to unsupported model name — falls back or errors

## How this service works

NoJunk LLM embeddings: OpenAI-compatible text embeddings on BGE-M3 (1024 dimensions, multilingual, up to 32 inputs per call).

## Output

Returns a JSON object conforming to the OpenAI embeddings response format: a list of embedding objects each containing the embedding index, a 1024-dimensional float array, and the object type 'embedding'; plus the model name ('bge-m3') and token usage stats (prompt_tokens and total_tokens).

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "input": {
   "type": [
    "string",
    "array"
   ],
   "description": "A string, or an array of 1-32 strings (each 1-8000 characters)"
  },
  "model": {
   "type": "string",
   "description": "Optional: bge-m3 (the only embedding model)"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "data": [
   {
    "index": 0,
    "object": "embedding",
    "embedding": [
     0.0213,
     -0.0471,
     0.0088,
     0.0329
    ]
   },
   {
    "index": 1,
    "object": "embedding",
    "embedding": [
     0.0213,
     -0.0471,
     0.0088,
     0.0329
    ]
   }
  ],
  "model": "bge-m3",
  "usage": {
   "total_tokens": 11,
   "prompt_tokens": 11
  },
  "object": "list"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/nojunk-llm-embeddings-bge-m3-5b622e75/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from nojunk-llm.mick8511.workers.dev](https://www.zero.xyz/host/nojunk-llm.mick8511.workers.dev/llms.txt)