# Koinelix Image Recognition

> Koinelix Image Recognition is a paid API for AI agents from api.koinelix.com, paid per call via x402, $0.011/call, status unknown (last checked 2026-11-16).

Analyzes an image (via URL or Base64) using a multimodal model and returns a text or Markdown description or answer to a user-supplied question about the image.

## Facts

- Endpoint: POST https://api.koinelix.com/v1/images/recognitions?utm_source=zero.xyz
- Price: $0.011/call
- Payment: x402
- Status: unknown
- Last checked: 2026-11-16
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/koinelix-image-recognition-e863d0a7
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_zo4FiQdnBZ2T1E7CfpA1v

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 koinelix-image-recognition-e863d0a7 -d '<json body>'
```

Example prompt: Can you look at this image — https://example.com/receipt.jpg — and tell me the total amount and date shown on the receipt? Give me the answer as plain text.

## When to prefer this

Choose this endpoint when an agent needs to understand, describe, or extract information from a single image without requiring an account or API key — especially for pay-per-call workflows using x402 micropayments. It is well-suited for reading visible text in photos, answering specific questions about image content, or generating descriptions for downstream tasks. Prefer it over general-purpose vision APIs when no account setup is desired and cost per call matters.

## Known failure modes

- Image URL is unreachable or returns a non-image content type — request fails with an error
- Base64 payload exceeds 10 MiB decoded — rejected with a size error
- Unsupported image format (not png, jpeg, webp, or gif) — rejected with a format error
- Prompt exceeds 4000 UTF-8 bytes — truncated or rejected
- Payment not provided or insufficient — HTTP 402 response blocking the call
- Both image_url and image_b64 provided simultaneously — ambiguous input error
- Network timeout fetching a remote image URL — transient failure

## How this service works

Image recognition and description with a multimodal model. Use when an agent needs to understand a picture: describe it, answer a question about it, read visible text or inspect details. One image goes in as an image URL or as Base64 (up to 10 MiB decoded) with an optional prompt; the answer returns as plain text or Markdown. Price: 0.011 USDC per call, paid with x402; no account or API key is needed.

## Output

A JSON object containing a unique recognition ID, the output format used (text or markdown), the model's answer as a string, a Unix timestamp of creation, and a boolean indicating whether the response was truncated due to length.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "prompt": {
   "type": "string",
   "minLength": 1,
   "description": "What to ask about the image, up to 4000 UTF-8 bytes. A default description request is used when omitted."
  },
  "image_b64": {
   "type": "string",
   "minLength": 1,
   "description": "The image as one standard padded Base64 string on a single line, without a data: prefix: png, jpeg, webp or gif, at most 10 MiB decoded. Produce it in code and put it into the request body (shell: base64 < photo.png | tr -d '\\n'; Python: base64.b64encode(data).decode()). Tip for clients that write tool-call arguments by hand, token by token: a long string is easily corrupted, so send a small image that way (a few thousand characters), or use image_url, or let code build the request. Use either image_url or image_b64.",
   "contentEncoding": "base64"
  },
  "image_url": {
   "type": "string",
   "format": "uri",
   "maxLength": 2048,
   "minLength": 1,
   "description": "Absolute http or https URL of the image. Preferred when the image is already online. Use either image_url or image_b64."
  },
  "output_format": {
   "enum": [
    "text",
    "markdown"
   ],
   "type": "string",
   "default": "text",
   "description": "Format of the answer."
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "id": "c_0123456789abcdef",
  "format": "markdown",
  "content": "A tabby cat sitting on a wooden windowsill.",
  "created": 1767225600,
  "truncated": false
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/koinelix-image-recognition-e863d0a7/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.koinelix.com](https://www.zero.xyz/host/api.koinelix.com/llms.txt)
