docs(local-models): add Rapid-MLX (Apple Silicon) provider page

[raullenchai]

Summary

Adds a new local-model docs page for Rapid-MLX,
an OpenAI-compatible inference server for Apple Silicon built on Apple's
MLX framework, with streaming and native tool calling.

Why

The current local-model pages cover Ollama, LM Studio, llamafile, and
Jan.ai, but none of them are MLX-native. Rapid-MLX fills this gap for
Apple Silicon users — it's pip-installable, exposes the standard
http://localhost:8000/v1 chat completions API, and works with the
existing --api_base flag, so no Open Interpreter code changes are
needed.

Files

• New: docs/language-models/local-models/rapid-mlx.mdx (modeled on
  lm-studio.mdx)
• Updated: docs/mint.json (registers the new page in the
  Local Providers nav group, between LM Studio and Custom Endpoint)

Scope

Docs only. No code, no API changes.