# Watt Multimodal Embeddings (Qwen3-VL)

> Watt Multimodal Embeddings (Qwen3-VL) is a paid API for AI agents from watt.area21.io, paid per call via x402, $0.002/call, status unknown (last checked 2026-10-21).

Generates 2048-dimensional embeddings for text and/or images using Qwen3-VL-Embedding-2B on Watt's own fleet, billed per call in USDC via x402.

## Facts

- Endpoint: POST https://watt.area21.io/v1/embeddings?utm_source=zero.xyz
- Price: $0.002/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-21
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/watt-multimodal-embeddings-qwen3-vl-69e49af5
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_Dgy_283SISHxfK093E0Ru

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 watt-multimodal-embeddings-qwen3-vl-69e49af5 -d '<json body>'
```

Example prompt: Embed this product description and its image together — the text is 'Vintage leather handbag, dark brown, handstitched' and the image is at https://example.com/bag.jpg — use 512-dimensional vectors and set the instruction to 'represent items for product similarity search'.

## When to prefer this

Choose this endpoint when you need multimodal embeddings that handle both text and images in a single vector — especially useful for vision-language retrieval, product search with images, or any RAG pipeline that indexes mixed content. Prefer it over text-only embedding APIs when your data includes images or image-text pairs. The OpenAI-compatible API format makes it a drop-in replacement with USDC pay-per-call pricing via x402, requiring no subscription.

## Known failure modes

- Missing required 'input' field returns a validation error
- Image URLs that are not publicly accessible will fail to fetch
- Batch exceeds 256 items limit returns an error
- Dimensions outside 64–2048 range are rejected
- Payment failure via x402 (insufficient USDC balance) blocks the request
- Unsupported model ID causes an error
- Invalid base64 or data URL for images causes a parsing error

## Output

Returns an OpenAI-compatible embeddings response containing one embedding object per input item. Each embedding is a float array (or base64 string) of the requested dimensionality (64–2048, default 2048), along with usage statistics (prompt tokens) and the model ID used.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method",
    "bodyType",
    "body"
   ],
   "properties": {
    "body": {
     "type": "object",
     "required": [
      "input"
     ],
     "properties": {
      "input": {
       "oneOf": [
        {
         "type": "string",
         "description": "A single text to embed."
        },
        {
         "type": "array",
         "items": {
          "oneOf": [
           {
            "type": "string",
            "description": "Text to embed."
           },
           {
            "type": "object",
            "properties": {
             "text": {
              "type": "string",
              "description": "Text paired with the image, or text alone."
             },
             "image": {
              "type": "string",
              "description": "Image as a data URL, base64, or public https URL. An x402 call is $0.002 when the request includes images, $0.001 for text-only."
             }
            },
            "description": "Text, an image, or both."
           }
          ],
          "description": "Text, an image, or both, embedded as one vector."
         },
         "maxItems": 256,
         "description": "Up to 256 items in one call."
        }
       ],
       "description": "One item, or a list of up to 256, batched 32 at a time. An item is text, an image, or both. An image is a data URL, base64, or a public https URL. An x402 call is $0.002 when the request includes images, $0.001 for text-only."
      },
      "model": {
       "type": "string",
       "description": "Embedding model id. qwen3-vl-embedding-2b when omitted."
      },
      "dimensions": {
       "type": "integer",
       "maximum": 2048,
       "minimum": 64,
       "description": "Matryoshka length of each vector, from 64 to 2048. Default 2048."
      },
      "instruction": {
       "type": "string",
       "description": "What the vectors are for. The model is instruction-aware; set this for retrieval."
      },
      "encoding_format": {
       "enum": [
        "float",
        "base64"
       ],
       "type": "string",
       "description": "float (default) or base64."
      }
     }
    },
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "POST",
      "PUT",
      "PATCH"
     ],
     "type": "string"
    }
… (truncated)
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/watt-multimodal-embeddings-qwen3-vl-69e49af5/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from watt.area21.io](https://www.zero.xyz/host/watt.area21.io/llms.txt)