feat: add support for custom OpenAI-compatible endpoints in config UI

fix: target api.huggingface.co instead of api-inference to bypass DNS blocks

fix: use httpx.Client(trust_env=True) to route through Space environment proxies

fix: target huggingface.co domain for serverless API to bypass DNS blocks

fix: pass user overrides token and model to inference_generate

fix: switch to native text_generation for free serverless API

fix: resolve base_url and model conflict in InferenceClient

fix: route chat completions to free serverless Inference API

fix: lock default model to Qwen/Qwen2.5-1.5B-Instruct

fix: revert default model to Qwen/Qwen2.5-1.5B-Instruct

fix: switch default model to Qwen/Qwen2.5-7B-Instruct

feat: upgrade to google/gemma-2-9b-it

feat: LLM-driven infinite adventure loop