# dev-5090 Vision / Image Understanding (Text-Fallback)

> dev-5090 Vision / Image Understanding (Text-Fallback) is a paid API for AI agents from dev-5090-box.tailfb4c1f.ts.net, paid per call via x402, $0.02/call, status unknown (last checked 2026-10-12).

Accepts image-understanding requests and currently returns text-fallback responses (no vision model loaded); billed at a reduced rate reflecting the preview/fallback status.

## Facts

- Endpoint: POST https://dev-5090-box.tailfb4c1f.ts.net/v1/vision?utm_source=zero.xyz
- Price: $0.02/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-12
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/dev-5090-vision-image-understanding-text-fallback-0c0dedbc
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_jSkGwcLsnXAp25GJfTDKH

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 dev-5090-vision-image-understanding-text-fallback-0c0dedbc -d '<json body>'
```

Example prompt: Can you tell me what's in this image? It's a photo from my last trip — just describe what you see in it, including any objects, text, or notable details.

## When to prefer this

Choose this endpoint when you need a low-cost, uncensored, preview-tier vision/image-understanding call and can tolerate text-fallback behavior (no real pixel analysis). Useful for testing integrations against an OpenAI-compatible vision endpoint, or when the text-fallback response is sufficient for your use case. Avoid if you require genuine image recognition or pixel-level analysis — wait for the vision model to be loaded.

## Known failure modes

- No vision model loaded — returns text-fallback response rather than genuine image analysis
- Payment required (402) if x402 header not included or USDC balance insufficient
- Malformed input schema (missing required fields) may return 400 or empty choices
- Network timeout on Tailscale-hosted endpoint if the device is offline
- Image URL inaccessible or unsupported format may result in fallback or empty response

## How this service works

[preview] Image understanding. Currently TEXT-FALLBACK ONLY — no vision model is loaded. Priced accordingly.

## Output

Returns an OpenAI-compatible JSON response with a 'choices' array containing the text-based response and a 'usage' object with token counts. Note: currently no vision model is loaded, so responses are text-fallback only — the model will not truly analyze pixel content.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "usage": {
   "type": "object"
  },
  "choices": {
   "type": "array"
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/dev-5090-vision-image-understanding-text-fallback-0c0dedbc/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from dev-5090-box.tailfb4c1f.ts.net](https://www.zero.xyz/host/dev-5090-box.tailfb4c1f.ts.net/llms.txt)