refactor(ui): 🔨 replace console.clear with cross-platform clear_screen function

Replaced all instances of `console.clear()` with a new `clear_screen()` helper function that uses native OS commands (`cls` for Windows, `clear` for Unix-like systems) instead of ANSI escape sequences.

- Adds `clear_screen()` function to launcher_tui.py, settings_tool.py, and credential_tool.py
- Replaces 18 instances of `console.clear()` across the codebase
- Improves terminal clearing reliability on classic Windows conhost and modern terminals (Windows Terminal, Linux, Mac)
- Removes unused anthropic_provider.py and bedrock_provider.py files
- Enhances credential_tool API key setup with better provider filtering logic to prevent duplicates
- Adds debug mode to show environment variable names in credential tool

src/proxy_app/launcher_tui.py

@@ -16,6 +16,17 @@ from dotenv import load_dotenv, set_key
 console = Console()
 
 
+def clear_screen():
+    """
+    Cross-platform terminal clear that works robustly on both 
+    classic Windows conhost and modern terminals (Windows Terminal, Linux, Mac).
+    
+    Uses native OS commands instead of ANSI escape sequences:
+    - Windows (conhost & Windows Terminal): cls
+    - Unix-like systems (Linux, Mac): clear
+    """
+    os.system('cls' if os.name == 'nt' else 'clear')
+
 class LauncherConfig:
     """Manages launcher_config.json (host, port, logging only)"""
     
@@ -262,7 +273,7 @@ class LauncherTUI:
     
     def show_main_menu(self):
         """Display main menu and handle selection"""
-        self.console.clear()
+        clear_screen()
         
         # Detect all settings
         settings = SettingsDetector.get_all_settings()
@@ -394,7 +405,7 @@ class LauncherTUI:
     def show_config_menu(self):
         """Display configuration sub-menu"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             self.console.print(Panel.fit(
                 "[bold cyan]⚙️  Proxy Configuration[/bold cyan]",
@@ -455,7 +466,7 @@ class LauncherTUI:
     
     def show_provider_settings_menu(self):
         """Display provider/advanced settings (read-only + launch tool)"""
-        self.console.clear()
+        clear_screen()
         
         settings = SettingsDetector.get_all_settings()
         credentials = settings["credentials"]
@@ -573,7 +584,7 @@ class LauncherTUI:
         import time
         
         # CRITICAL: Show full loading UI to replace the 6-7 second blank wait
-        self.console.clear()
+        clear_screen()
         
         _start_time = time.time()
         
@@ -610,7 +621,7 @@ class LauncherTUI:
     
     def show_about(self):
         """Display About page with project information"""
-        self.console.clear()
+        clear_screen()
         
         self.console.print(Panel.fit(
             "[bold cyan]ℹ️  About LLM API Key Proxy[/bold cyan]",
@@ -654,7 +665,7 @@ class LauncherTUI:
         """Prepare and launch proxy in same window"""
         # Check if forced onboarding needed
         if self.needs_onboarding():
-            self.console.clear()
+            clear_screen()
             self.console.print(Panel(
                 Text.from_markup(
                     "⚠️  [bold yellow]Setup Required[/bold yellow]\n\n"
@@ -677,13 +688,13 @@ class LauncherTUI:
                 return
         
         # Clear console and modify sys.argv
-        self.console.clear()
+        clear_screen()
         self.console.print(f"\n[bold green]🚀 Starting proxy on {self.config.config['host']}:{self.config.config['port']}...[/bold green]\n")
         
         # Clear console again to remove the starting message before main.py shows loading details
         import time
         time.sleep(0.5)  # Brief pause so user sees the message
-        self.console.clear()
+        clear_screen()
         
         # Reconstruct sys.argv for main.py
         sys.argv = [

src/proxy_app/main.py

@@ -1137,7 +1137,7 @@ if __name__ == "__main__":
 
     def show_onboarding_message():
         """Display clear explanatory message for why onboarding is needed."""
-        console.clear()  # Clear terminal for clean presentation
+        os.system('cls' if os.name == 'nt' else 'clear')  # Clear terminal for clean presentation
         console.print(Panel.fit(
             "[bold cyan]🚀 LLM API Key Proxy - First Time Setup[/bold cyan]",
             border_style="cyan"

src/proxy_app/settings_tool.py

@@ -15,6 +15,18 @@ from dotenv import set_key, unset_key
 console = Console()
 
 
+def clear_screen():
+    """
+    Cross-platform terminal clear that works robustly on both 
+    classic Windows conhost and modern terminals (Windows Terminal, Linux, Mac).
+    
+    Uses native OS commands instead of ANSI escape sequences:
+    - Windows (conhost & Windows Terminal): cls
+    - Unix-like systems (Linux, Mac): clear
+    """
+    os.system('cls' if os.name == 'nt' else 'clear')
+
+
 class AdvancedSettings:
     """Manages pending changes to .env"""
     
@@ -389,7 +401,7 @@ class SettingsTool:
     
     def show_main_menu(self):
         """Display settings categories"""
-        self.console.clear()
+        clear_screen()
         
         self.console.print(Panel.fit(
             "[bold cyan]🔧 Advanced Settings Configuration[/bold cyan]",
@@ -436,7 +448,7 @@ class SettingsTool:
     def manage_custom_providers(self):
         """Manage custom provider API bases"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             providers = self.provider_mgr.get_current_providers()
             
@@ -533,7 +545,7 @@ class SettingsTool:
     def manage_model_definitions(self):
         """Manage provider model definitions"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             all_providers = self.model_mgr.get_all_providers_with_models()
             
@@ -710,7 +722,7 @@ class SettingsTool:
             current_models = {m: {} for m in current_models}
         
         while True:
-            self.console.clear()
+            clear_screen()
             self.console.print(f"[bold]Editing models for: {provider}[/bold]\n")
             self.console.print("Current models:")
             for i, (name, definition) in enumerate(current_models.items(), 1):
@@ -788,7 +800,7 @@ class SettingsTool:
             input("\nPress Enter to continue...")
             return
         
-        self.console.clear()
+        clear_screen()
         self.console.print(f"[bold]Provider: {provider}[/bold]\n")
         self.console.print("[bold]📦 Configured Models:[/bold]")
         self.console.print("━" * 50)
@@ -816,7 +828,7 @@ class SettingsTool:
     def manage_provider_settings(self):
         """Manage provider-specific settings (Antigravity, Gemini CLI)"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             available_providers = self.provider_settings_mgr.get_available_providers()
             
@@ -863,7 +875,7 @@ class SettingsTool:
     def _manage_single_provider_settings(self, provider: str):
         """Manage settings for a single provider"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             display_name = provider.replace("_", " ").title()
             definitions = self.provider_settings_mgr.get_provider_settings_definitions(provider)
@@ -1005,7 +1017,7 @@ class SettingsTool:
     def manage_concurrency_limits(self):
         """Manage concurrency limits"""
         while True:
-            self.console.clear()
+            clear_screen()
             
             limits = self.concurrency_mgr.get_current_limits()
             

src/rotator_library/credential_tool.py

@@ -37,6 +37,18 @@ def _ensure_providers_loaded():
     return _provider_factory, _provider_plugins
 
 
+def clear_screen():
+    """
+    Cross-platform terminal clear that works robustly on both 
+    classic Windows conhost and modern terminals (Windows Terminal, Linux, Mac).
+    
+    Uses native OS commands instead of ANSI escape sequences:
+    - Windows (conhost & Windows Terminal): cls
+    - Unix-like systems (Linux, Mac): clear
+    """
+    os.system('cls' if os.name == 'nt' else 'clear')
+
+
 def _get_credential_number_from_filename(filename: str) -> int:
     """
     Extract credential number from filename like 'provider_oauth_1.json' -> 1
@@ -127,6 +139,9 @@ async def setup_api_key():
     """
     console.print(Panel("[bold cyan]API Key Setup[/bold cyan]", expand=False))
 
+    # Debug toggle: Set to True to see env var names next to each provider
+    SHOW_ENV_VAR_NAMES = True
+
     # Verified list of LiteLLM providers with their friendly names and API key variables
     LITELLM_PROVIDERS = {
         "OpenAI": "OPENAI_API_KEY", "Anthropic": "ANTHROPIC_API_KEY",
@@ -162,26 +177,59 @@ async def setup_api_key():
         "Nscale": "NSCALE_API_KEY", "Recraft": "RECRAFT_API_KEY",
         "v0": "V0_API_KEY", "Vercel": "VERCEL_AI_GATEWAY_API_KEY",
         "Topaz": "TOPAZ_API_KEY", "ElevenLabs": "ELEVENLABS_API_KEY",
-        "Deepgram": "DEEPGRAM_API_KEY", "Custom API": "CUSTOM_API_KEY",
+        "Deepgram": "DEEPGRAM_API_KEY",
         "GitHub Models": "GITHUB_TOKEN", "GitHub Copilot": "GITHUB_COPILOT_API_KEY",
     }
 
     # Discover custom providers and add them to the list
-    # Note: gemini_cli is OAuth-only, but qwen_code and iflow support both OAuth and API keys
+    # Note: gemini_cli and antigravity are OAuth-only
+    # qwen_code API key support is a fallback
+    # iflow API key support is a feature
     _, PROVIDER_PLUGINS = _ensure_providers_loaded()
-    oauth_only_providers = {'gemini_cli', 'antigravity'}
-    discovered_providers = {
-        p.replace('_', ' ').title(): p.upper() + "_API_KEY"
-        for p in PROVIDER_PLUGINS.keys()
-        if p not in oauth_only_providers and p.replace('_', ' ').title() not in LITELLM_PROVIDERS
+    
+    # Build a set of environment variables already in LITELLM_PROVIDERS
+    # to avoid duplicates based on the actual API key names
+    litellm_env_vars = set(LITELLM_PROVIDERS.values())
+    
+    # Providers to exclude from API key list
+    exclude_providers = {
+        'gemini_cli',  # OAuth-only
+        'antigravity',  # OAuth-only  
+        'qwen_code',  # API key is fallback, OAuth is primary - don't advertise
+        'openai_compatible',  # Base class, not a real provider
     }
     
+    discovered_providers = {}
+    for provider_key in PROVIDER_PLUGINS.keys():
+        if provider_key in exclude_providers:
+            continue
+        
+        # Create environment variable name
+        env_var = provider_key.upper() + "_API_KEY"
+        
+        # Check if this env var already exists in LITELLM_PROVIDERS
+        # This catches duplicates like GEMINI_API_KEY, MISTRAL_API_KEY, etc.
+        if env_var in litellm_env_vars:
+            # Already in LITELLM_PROVIDERS with better name, skip this one
+            continue
+        
+        # Create display name for this custom provider
+        display_name = provider_key.replace('_', ' ').title()
+        discovered_providers[display_name] = env_var
+    
+    # LITELLM_PROVIDERS takes precedence (comes first in merge)
     combined_providers = {**LITELLM_PROVIDERS, **discovered_providers}
     provider_display_list = sorted(combined_providers.keys())
 
     provider_text = Text()
     for i, provider_name in enumerate(provider_display_list):
-        provider_text.append(f"  {i + 1}. {provider_name}\n")
+        if SHOW_ENV_VAR_NAMES:
+            # Extract env var prefix (before _API_KEY)
+            env_var = combined_providers[provider_name]
+            prefix = env_var.replace("_API_KEY", "").replace("_", " ")
+            provider_text.append(f"  {i + 1}. {provider_name} ({prefix})\n")
+        else:
+            provider_text.append(f"  {i + 1}. {provider_name}\n")
 
     console.print(Panel(provider_text, title="Available Providers for API Key", style="bold blue"))
 
@@ -1000,7 +1048,7 @@ async def export_credentials_submenu():
     Submenu for credential export options.
     """
     while True:
-        console.clear()
+        clear_screen()
         console.print(Panel("[bold cyan]Export Credentials to .env[/bold cyan]", title="--- API Key Proxy ---", expand=False))
         
         console.print(Panel(
@@ -1111,7 +1159,7 @@ async def main(clear_on_start=True):
     
     while True:
         # Clear screen between menu selections for cleaner UX
-        console.clear()
+        clear_screen()
         console.print(Panel("[bold cyan]Interactive Credential Setup[/bold cyan]", title="--- API Key Proxy ---", expand=False))
         
         console.print(Panel(
@@ -1179,7 +1227,7 @@ async def main(clear_on_start=True):
         elif setup_type == "2":
             await setup_api_key()
             #console.print("\n[dim]Press Enter to return to main menu...[/dim]")
-            input()
+            #input()
 
         elif setup_type == "3":
             await export_credentials_submenu()
@@ -1225,7 +1273,7 @@ def run_credential_tool(from_launcher=False):
     # If from launcher, don't clear screen at start to preserve loading messages
     try:
         asyncio.run(main(clear_on_start=not from_launcher))
-        console.clear()  # Clear terminal when credential tool exits
+        clear_screen()  # Clear terminal when credential tool exits
     except KeyboardInterrupt:
         console.print("\n[bold yellow]Exiting setup.[/bold yellow]")
-        console.clear()  # Clear terminal on keyboard interrupt too
+        clear_screen()  # Clear terminal on keyboard interrupt too

src/rotator_library/providers/anthropic_provider.py

@@ -1,31 +0,0 @@
-import httpx
-import logging
-from typing import List
-from .provider_interface import ProviderInterface
-
-lib_logger = logging.getLogger('rotator_library')
-lib_logger.propagate = False # Ensure this logger doesn't propagate to root
-if not lib_logger.handlers:
-    lib_logger.addHandler(logging.NullHandler())
-
-class AnthropicProvider(ProviderInterface):
-    """
-    Provider implementation for the Anthropic API.
-    """
-    async def get_models(self, api_key: str, client: httpx.AsyncClient) -> List[str]:
-        """
-        Fetches the list of available models from the Anthropic API.
-        """
-        try:
-            response = await client.get(
-                "https://api.anthropic.com/v1/models",
-                headers={
-                    "x-api-key": api_key,
-                    "anthropic-version": "2023-06-01"
-                }
-            )
-            response.raise_for_status()
-            return [f"anthropic/{model['id']}" for model in response.json().get("data", [])]
-        except httpx.RequestError as e:
-            lib_logger.error(f"Failed to fetch Anthropic models: {e}")
-            return []

src/rotator_library/providers/bedrock_provider.py

@@ -1,29 +0,0 @@
-import httpx
-import logging
-from typing import List
-from .provider_interface import ProviderInterface
-
-lib_logger = logging.getLogger('rotator_library')
-lib_logger.propagate = False # Ensure this logger doesn't propagate to root
-if not lib_logger.handlers:
-    lib_logger.addHandler(logging.NullHandler())
-
-class BedrockProvider(ProviderInterface):
-    """
-    Provider implementation for AWS Bedrock.
-    """
-    async def get_models(self, api_key: str, client: httpx.AsyncClient) -> List[str]:
-        """
-        Returns a hardcoded list of common Bedrock models, as there is no
-        simple, unauthenticated API endpoint to list them.
-        """
-        # Note: Listing Bedrock models typically requires AWS credentials and boto3.
-        # For a simple, key-based proxy, we'll list common models.
-        # This can be expanded with full AWS authentication if needed.
-        lib_logger.info("Returning hardcoded list for Bedrock. Full discovery requires AWS auth.")
-        return [
-            "bedrock/anthropic.claude-3-sonnet-20240229-v1:0",
-            "bedrock/anthropic.claude-3-haiku-20240307-v1:0",
-            "bedrock/cohere.command-r-plus-v1:0",
-            "bedrock/mistral.mistral-large-2402-v1:0",
-        ]