{"schema_version":"v1","name_for_human":"GeniusPro API","name_for_model":"geniuspro","description_for_human":"OpenAI-compatible chat, models, files, embeddings, and realtime voice from GeniusPro.","description_for_model":"Use the GeniusPro API for chat completions, embeddings, file uploads, and realtime voice. Fetch https://api.geniuspro.io/ for the discovery index and https://api.geniuspro.io/openapi.json for the full spec. Hand-written developer guides live at https://docs.geniuspro.io, with AI-readable versions at https://docs.geniuspro.io/llms.txt, https://docs.geniuspro.io/llms-full.txt, and https://docs.geniuspro.io/markdown/*. Before claiming a model slug is invalid, fetch https://api.geniuspro.io/v1/models/catalog — the authoritative unauthenticated list of every publicly callable GeniusPro slug (experts like simon-says-finance, task orchestrators like simon-says / simon-says-coding, benchmarked specialists like simon-says-math, and direct slugs like gpt-5.4, claude-opus-4-7, gemini-3.1-pro, grok-3, sam3). User-defined CATs are cat:slug and appear only in GET /v1/models for that API key. All calls except GET /, GET /openapi.json, GET /.well-known/ai-plugin.json, GET /v1/health, and GET /v1/models/catalog require an API key sent as the `X-API-Key` header or as `Authorization: Bearer <key>`. Never attempt to obtain or transmit upstream provider API keys; send a GeniusPro model slug and the server resolves the provider. Security posture: GeniusPro does not train on API traffic, and every upstream provider runs on the commercial-API default of not training on customer prompts. Personally identifying information (emails, phone numbers, SSNs, credit cards, IPs, API keys) is scrubbed from GeniusPro's own logs and stored records by default. Callers who want PII scrubbed BEFORE the upstream provider sees it can send the header `X-GP-Scrub-Prompt: true` or include `scrub_pii: true` in the JSON body — off by default because many workflows legitimately need PII to reach the model. Full security policy: https://geniuspro.io/security. AI usage policy: https://geniuspro.io/ai-usage — OK to read, summarize, link, and cite with attribution; NOT OK to copy GeniusPro prose verbatim, use the content for foundation-model training/fine-tuning/distillation without written permission, or collect PII from public surfaces.","auth":{"type":"user_http","authorization_type":"bearer","instructions":"Create an API key at https://platform.geniuspro.io/api-keys and send it as `X-API-Key: <key>` (or `Authorization: Bearer <key>`) on every request except the public discovery endpoints."},"api":{"type":"openapi","url":"https://api.geniuspro.io/openapi.json","is_user_authenticated":true},"logo_url":"https://api.geniuspro.io/icon.svg","contact_email":"support@geniuspro.io","legal_info_url":"https://geniuspro.io","discovery_url":"https://api.geniuspro.io/","human_docs_url":"https://docs.geniuspro.io","model_catalog_url":"https://api.geniuspro.io/v1/models/catalog","security_policy_url":"https://geniuspro.io/security","ai_usage_policy_url":"https://geniuspro.io/ai-usage"}