feat: add Chinese UI, Responses models, and mirrored Windows setup

Localize the web console and persist custom model and capability settings.
Use OpenAI Responses exclusively and move runtime limits into environment configuration.
Bootstrap uv and the virtual environment on Windows, support optional China mirrors,
and package source distributions without environments, build output, or credentials.

Update regression tests and repository quality checks for the new behavior.

artemis/memory/chunking.py

@@ -35,6 +35,7 @@ context is rebuilt only when a compression event replaces or folds turns.
 """
 
 import asyncio
+from artemis.config.settings import settings
 from dataclasses import dataclass, field
 import json
 from pathlib import Path
@@ -45,7 +46,6 @@ from uuid import uuid4
 
 from langchain_core.messages import BaseMessage, HumanMessage, SystemMessage
 
-from artemis.llm.google import is_google_provider
 from artemis.memory.step_memory import JobKey, StepLens, StepMemoryService
 from artemis.memory.transcript import format_session_offset
 from artemis.utils.logger import get_logger
@@ -298,7 +298,7 @@ class StepCapsuleLens(StepLens):
         fallback_model_name: str | None = None,
         fallback_llm: Any | None = None,
     ):
-        self._model_name = model_name or "gemini-3.8-flash"
+        self._model_name = model_name or settings.OPENAI_MODEL
         self._llm = llm
         self._ctx = ctx
         # Availability hardening: `chunking.model` is a dedicated model with no
@@ -323,16 +323,16 @@ class StepCapsuleLens(StepLens):
 
     def _get_llm(self):
         if self._llm is None:
-            from artemis.services.llm import get_google_llm
+            from artemis.services.llm import get_openai_llm
 
-            self._llm = get_google_llm(model_name=self._model_name, temperature=0.0)
+            self._llm = get_openai_llm(model_name=self._model_name, temperature=0.0)
         return self._llm
 
     def _get_fallback_llm(self):
         if self._fallback_llm is None and self._fallback_model_name:
-            from artemis.services.llm import get_google_llm
+            from artemis.services.llm import get_openai_llm
 
-            self._fallback_llm = get_google_llm(
+            self._fallback_llm = get_openai_llm(
                 model_name=self._fallback_model_name, temperature=0.0
             )
         return self._fallback_llm
@@ -891,7 +891,7 @@ class HistoryChunkManager:
         self._max_steps = int(getattr(cc, "max_steps", 12) or 12)
         self._min_steps = max(1, min(self._max_steps, int(getattr(cc, "min_steps", 3) or 3)))
         self._target_source_tokens = int(getattr(cc, "target_source_tokens", 2000) or 2000)
-        self._model_name = getattr(cc, "model", None) or "gemini-3.8-flash"
+        self._model_name = getattr(cc, "model", None) or settings.OPENAI_MODEL
         self._max_chunks = int(getattr(cc, "max_chunks", 8) or 8)
         # None uses max_chunks as the era cap.
         self._max_eras = int(getattr(cc, "max_eras", None) or self._max_chunks)
@@ -965,7 +965,7 @@ class HistoryChunkManager:
             fallback = getattr(getattr(llm_cfg, "summarizer", None), "fallback", None)
             provider = str(getattr(fallback, "provider", "") or "")
             model = getattr(fallback, "model", None)
-            if model and is_google_provider(provider) and model != self._model_name:
+            if model and str(provider).lower() == "openai" and model != self._model_name:
                 return str(model)
         except Exception as exc:
             logger.debug(f"Capsule fallback model resolution skipped: {exc}", exc_info=True)