refactor: remove OpenRouter rollback, shims, and redundant layers

- OpenRouter: native Anthropic only; remove chat_request and OPENROUTER_TRANSPORT
- Drop OpenAICompatibleProvider alias, api.request_utils, voice_pipeline facade
- Simplify OpenRouter SSE, generic reasoning in conversion, messaging dispatch
- Shared markdown table helpers; API optimization response helper; contract guards
- Restore PLAN.md; update docs and tests

api/app.py

@@ -13,7 +13,7 @@ from providers.exceptions import ProviderError
 
 from .dependencies import cleanup_provider
 from .routes import router
-from .runtime import AppRuntime, warn_if_process_auth_token
+from .runtime import AppRuntime
 
 # Opt-in to future behavior for python-telegram-bot
 os.environ["PTB_TIMEDELTA"] = "1"
@@ -23,11 +23,6 @@ _settings = get_settings()
 configure_logging(_settings.log_file)
 
 
-def _warn_if_process_auth_token(settings) -> None:
-    """Compatibility wrapper for tests importing the old app helper."""
-    warn_if_process_auth_token(settings)
-
-
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     """Application lifespan manager."""

api/optimization_handlers.py

@@ -22,6 +22,22 @@ from .models.anthropic import MessagesRequest
 from .models.responses import MessagesResponse, Usage
 
 
+def _text_response(
+    request_data: MessagesRequest,
+    text: str,
+    *,
+    input_tokens: int,
+    output_tokens: int,
+) -> MessagesResponse:
+    return MessagesResponse(
+        id=f"msg_{uuid.uuid4()}",
+        model=request_data.model,
+        content=[{"type": "text", "text": text}],
+        stop_reason="end_turn",
+        usage=Usage(input_tokens=input_tokens, output_tokens=output_tokens),
+    )
+
+
 def try_prefix_detection(
     request_data: MessagesRequest, settings: Settings
 ) -> MessagesResponse | None:
@@ -34,12 +50,11 @@ def try_prefix_detection(
         return None
 
     logger.info("Optimization: Fast prefix detection request")
-    return MessagesResponse(
-        id=f"msg_{uuid.uuid4()}",
-        model=request_data.model,
-        content=[{"type": "text", "text": extract_command_prefix(command)}],
-        stop_reason="end_turn",
-        usage=Usage(input_tokens=100, output_tokens=5),
+    return _text_response(
+        request_data,
+        extract_command_prefix(command),
+        input_tokens=100,
+        output_tokens=5,
     )
 
 
@@ -53,13 +68,11 @@ def try_quota_mock(
         return None
 
     logger.info("Optimization: Intercepted and mocked quota probe")
-    return MessagesResponse(
-        id=f"msg_{uuid.uuid4()}",
-        model=request_data.model,
-        role="assistant",
-        content=[{"type": "text", "text": "Quota check passed."}],
-        stop_reason="end_turn",
-        usage=Usage(input_tokens=10, output_tokens=5),
+    return _text_response(
+        request_data,
+        "Quota check passed.",
+        input_tokens=10,
+        output_tokens=5,
     )
 
 
@@ -73,13 +86,11 @@ def try_title_skip(
         return None
 
     logger.info("Optimization: Skipped title generation request")
-    return MessagesResponse(
-        id=f"msg_{uuid.uuid4()}",
-        model=request_data.model,
-        role="assistant",
-        content=[{"type": "text", "text": "Conversation"}],
-        stop_reason="end_turn",
-        usage=Usage(input_tokens=100, output_tokens=5),
+    return _text_response(
+        request_data,
+        "Conversation",
+        input_tokens=100,
+        output_tokens=5,
     )
 
 
@@ -93,13 +104,11 @@ def try_suggestion_skip(
         return None
 
     logger.info("Optimization: Skipped suggestion mode request")
-    return MessagesResponse(
-        id=f"msg_{uuid.uuid4()}",
-        model=request_data.model,
-        role="assistant",
-        content=[{"type": "text", "text": ""}],
-        stop_reason="end_turn",
-        usage=Usage(input_tokens=100, output_tokens=1),
+    return _text_response(
+        request_data,
+        "",
+        input_tokens=100,
+        output_tokens=1,
     )
 
 
@@ -116,13 +125,11 @@ def try_filepath_mock(
 
     filepaths = extract_filepaths_from_command(cmd, output)
     logger.info("Optimization: Mocked filepath extraction")
-    return MessagesResponse(
-        id=f"msg_{uuid.uuid4()}",
-        model=request_data.model,
-        role="assistant",
-        content=[{"type": "text", "text": filepaths}],
-        stop_reason="end_turn",
-        usage=Usage(input_tokens=100, output_tokens=10),
+    return _text_response(
+        request_data,
+        filepaths,
+        input_tokens=100,
+        output_tokens=10,
     )
 
 

api/request_utils.py

@@ -1,5 +0,0 @@
-"""Backward-compatible token counting import for API route handlers."""
-
-from core.anthropic import get_token_count
-
-__all__ = ["get_token_count"]

api/routes.py

@@ -4,11 +4,11 @@ from fastapi import APIRouter, Depends, HTTPException, Request, Response
 from loguru import logger
 
 from config.settings import Settings
+from core.anthropic import get_token_count
 
 from .dependencies import get_provider_for_type, get_settings, require_api_key
 from .models.anthropic import MessagesRequest, TokenCountRequest
 from .models.responses import ModelResponse, ModelsListResponse
-from .request_utils import get_token_count
 from .services import ClaudeProxyService
 
 router = APIRouter()
@@ -75,7 +75,6 @@ def _probe_response(allow: str) -> Response:
 @router.post("/v1/messages")
 async def create_message(
     request_data: MessagesRequest,
-    _raw_request: Request,
     service: ClaudeProxyService = Depends(get_proxy_service),
     _auth=Depends(require_api_key),
 ):

api/services.py

@@ -12,7 +12,7 @@ from fastapi.responses import StreamingResponse
 from loguru import logger
 
 from config.settings import Settings
-from core.anthropic import get_user_facing_error_message
+from core.anthropic import get_token_count, get_user_facing_error_message
 from providers.base import BaseProvider
 from providers.exceptions import InvalidRequestError, ProviderError
 
@@ -20,7 +20,6 @@ from .model_router import ModelRouter
 from .models.anthropic import MessagesRequest, TokenCountRequest
 from .models.responses import TokenCountResponse
 from .optimization_handlers import try_optimizations
-from .request_utils import get_token_count
 
 TokenCounter = Callable[[list[Any], str | list[Any] | None, list[Any] | None], int]