Per-model docs are live: 48 models, 4 protocols, one Base URL

Every model now has its own page with endpoints, auth headers and copy-ready curl, Python and Node samples. One gateway, four protocols, no per-vendor SDK juggling.

• Coverage: 48 representative models across chat, image, video and audio.
• Four protocols, one Base URL: OpenAI, Anthropic, Gemini and media tasks.
• Every model can also be opened directly in the workspace, no code needed.