generated: '2026-08-11' method: derived status: candidate source: openapi/anyimagedetector-ai-image-detector-openapi.yml description: AnyImageDetector operates NO MCP server — hosted, remote or local — and publishes no agent surface of any kind beyond an llms.txt that describes marketing pages. The tool below is a CANDIDATE derived from the single documented REST operation, recorded so the shape of a future server is on file. It is not deployed by the provider and no `MCPServer` pointer is wired into apis.yml, because asserting one would credit an agent surface that does not exist. search_result: official_hosted_server: none local_server_package: none registry_listings: none probes: - url: https://imagedetector.online/mcp method: POST tools/list status: 500 body: '{"error":"Only HTML requests are supported here"}' - url: https://imagedetector.online/v1/mcp method: POST tools/list status: 500 body: '{"error":"Only HTML requests are supported here"}' - url: https://imagedetector.online/api/mcp method: POST tools/list status: 500 body: '{"error":"Only HTML requests are supported here"}' - url: https://imagedetector.online/sse method: POST tools/list status: 500 body: '{"error":"Only HTML requests are supported here"}' - url: https://imagedetector.online/.well-known/oauth-authorization-server method: GET status: 404 - registry: npm query: anyimagedetector / imagedetector result: no first-party MCP package note: The 500 "Only HTML requests are supported here" is the Next.js application rejecting a non-HTML request on an unrouted path — it is a negative result, not a server. candidate_tools: - name: detect_image title: Detect whether an image is AI-generated description: Analyze an image and return a verdict (likely_ai | likely_human | uncertain), an ai_score from 0 to 1, a confidence label (low | medium | high), and a source_breakdown array. Consumes 1 credit per successful detection. rest_operation_id: detectImage rest: POST /v1/image/detect input_schema_source: Inherited verbatim from the requestBody of detectImage in openapi/anyimagedetector-ai-image-detector-openapi.yml — exactly one of `image` (multipart file, JPG/PNG/WebP, <= 8 MB) or `imageUrl` (public HTTP/HTTPS URL). action_class: acting consequence: write reason_not_read_only: The call is metered and billable — each success debits a credit — so it is not a free read even though it mutates no provider-side resource the caller owns. auth: static API key (Bearer sk_... or x-api-key), paid plan required agent_readiness_notes: - No MCP server, no A2A agent card, no agent skills published by the provider. - robots.txt explicitly Disallows ClaudeBot, GPTBot, Google-Extended, CCBot, Bytespider, Amazonbot, Applebot-Extended and meta-externalagent, and sets Cloudflare Content Signals to ai-train=no / use=reference — the provider's stated posture toward AI agents on the website is restrictive, while the API itself is sold as a machine surface. deployment: mode: none verified: derived checked: '2026-08-12' source: catalog MCP census