Summary

Add an OpenAI-backed text generation provider alongside the existing Ollama flow.

Proposed changes

• add provider-aware config fields for OpenAI and Ollama
• teach the LLM adapter to call the OpenAI Responses API
• update startup/model selection and cron flow for provider-aware execution
• document the new config and environment-variable setup

Why

This makes MPV2 usable with a hosted model provider and improves output quality for users who do not want to rely on a local Ollama model.