Tools / AI
Image Description $0.005 / call
Describe any public image, including its text.
GET /v1/describe-imageFetches a public image (up to 3MB) and returns an AI description including any visible text, using a vision model. Useful for alt-text, OCR-lite, and image understanding in agent workflows.
Parameters
url required | Public image URL (max 3MB) |
prompt | Optional custom question about the image |
Example
macOS / Linuxcurl -s "https://quartermaster.surewhynot.app/v1/describe-image?url=https://www.google.com/images/branding/googlelogo/1x/googlelogo_color_272x92dp.png"curl.exe -s "https://quartermaster.surewhynot.app/v1/describe-image?url=https://www.google.com/images/branding/googlelogo/1x/googlelogo_color_272x92dp.png"Unpaid, these return 402 Payment Required with the payment details in the header (add -i to see them). curl.exe ships with Windows 10+ and works in cmd and both PowerShell versions. To actually pay, use an x402 client (see Connect) or call it as the MCP tool. Cost: $0.005 per call.
FAQ
Can it read text in the image?
Yes — the description includes visible text, so it doubles as lightweight OCR.
How do I pay for this endpoint?
Call it with no payment and you receive HTTP 402 with machine-readable payment requirements. Attach a USDC payment on Base with any x402 client (for example @x402/fetch) and retry — the whole loop is two requests. This call costs $0.005 in USDC.
Do I need an account or API key?
No. There are no accounts, logins, or API keys anywhere in Quartermaster. Payment is per call over the x402 protocol; your wallet is your identity.
Can I call this as an MCP tool instead?
Yes. Quartermaster is also an MCP server at /mcp, so any MCP client (Claude Code, Claude.ai, Cursor) can mount this tool. See /connect for setup.