@oneciel-ai/claude-any 0.1.37 → 0.1.39

package/README.md

@@ -48,7 +48,7 @@ arguments through unchanged.
 
 Credits: One Ciel LLC
 
-Current version: `0.1.37`
+Current version: `0.1.39`
 
 ## Why This Exists
 
@@ -381,6 +381,23 @@ steps under that larger model's supervision.
 
 ## Changelog
 
+### 0.1.39
+
+- **Menu input fixes**: restores terminal line/echo mode before text or number
+  prompts, so typed numeric values are visible in the prelaunch UI.
+- **Safer numeric validation**: invalid numeric option input now shows an
+  inline message instead of crashing the menu.
+- **Preset visibility**: applied presets report the effective context, reserve,
+  output, and timeout values.
+
+### 0.1.38
+
+- **User-selected context windows**: removes the NVIDIA hosted 32K safety cap.
+  The router now uses the context window selected in LLM options or headless
+  configuration, with model-aware fallback only when no value is configured.
+- **NVIDIA presets updated**: NVIDIA hosted presets now start at 65K and scale
+  up to 256K for large-output/reasoning workflows.
+
 ### 0.1.37
 
 - **Pseudo tool-call recovery**: the NVIDIA/OpenAI-compatible stream path now

package/claude_any.py

@@ -85,7 +85,7 @@ PROVIDER_LABELS = {
     "self-hosted-nim": "Self Hosted NIM",
 }
 APP_NAME = "Claude Any"
-VERSION = "0.1.37"
+VERSION = "0.1.39"
 CREDITS = "Credits: One Ciel LLC"
 
 LOG_LEVELS = {"SILENT": 0, "ERROR": 1, "WARN": 2, "INFO": 3, "DEBUG": 4, "TRACE": 5}
@@ -144,7 +144,7 @@ LANGUAGES = {
     "zh": "中文",
 }
 
-MODEL_PRESETS: dict[str, dict[str, Any]] = {
+MODEL_PRESETS: dict[str, dict[str, Any]] = {
     "glm-4.7": {"compat_max_tokens": 64, "thinking": True, "num_ctx_min": 32768, "num_ctx_max": 131072},
     "glm-5.1": {"compat_max_tokens": 64, "thinking": True, "num_ctx_min": 32768, "num_ctx_max": 131072},
     "glm-4.7:cloud": {"compat_max_tokens": 64, "thinking": True, "num_ctx_min": 32768, "num_ctx_max": 131072},
@@ -154,10 +154,21 @@ MODEL_PRESETS: dict[str, dict[str, Any]] = {
     "qwen3.6:27b": {"compat_max_tokens": 16, "thinking": False, "num_ctx_min": 32768, "num_ctx_max": 65536},
     "deepseek-r1": {"compat_max_tokens": 64, "thinking": True, "num_ctx_min": 32768, "num_ctx_max": 131072},
     "llama3.3:70b": {"compat_max_tokens": 16, "thinking": False, "num_ctx_min": 32768, "num_ctx_max": 131072},
-}
-
-
-def model_preset(model_id: str) -> dict[str, Any]:
+}
+
+
+def nvidia_hosted_context_default(model_id: str) -> int:
+    model = model_id.lower()
+    if "kimi-k2.6" in model or "kimi_k2.6" in model:
+        return 262144
+    if "deepseek" in model:
+        return 131072
+    if "glm" in model or "qwen" in model:
+        return 65536
+    return 65536
+
+
+def model_preset(model_id: str) -> dict[str, Any]:
     """Return preset dict for a model ID, checking exact match then prefix match."""
     if model_id in MODEL_PRESETS:
         return MODEL_PRESETS[model_id]
@@ -722,7 +733,7 @@ DEFAULT_CONFIG: dict[str, Any] = {
             "native_compat": False,
             "rate_limit_rpm": 40,
             "rate_limit_status": True,
-            "context_window": 32768,
+            "context_window": 65536,
             "max_output_tokens": 4096,
             "temperature": 0.7,
             "top_p": 0.8,
@@ -788,7 +799,14 @@ def apply_config_migrations(cfg: dict[str, Any]) -> None:
     if not migrations.get(marker):
         pcfg = cfg.get("providers", {}).get("nvidia-hosted", {})
         if isinstance(pcfg, dict) and not positive_int(pcfg.get("context_window")):
-            pcfg["context_window"] = 32768
+            pcfg["context_window"] = nvidia_hosted_context_default(str(pcfg.get("current_model") or ""))
+        migrations[marker] = True
+
+    marker = "nvidia_context_window_unforce_32k_20260513"
+    if not migrations.get(marker):
+        pcfg = cfg.get("providers", {}).get("nvidia-hosted", {})
+        if isinstance(pcfg, dict) and positive_int(pcfg.get("context_window")) == 32768:
+            pcfg["context_window"] = nvidia_hosted_context_default(str(pcfg.get("current_model") or ""))
         migrations[marker] = True
 
     marker = "stream_enabled_default_true_20260513"
@@ -3620,7 +3638,7 @@ def openai_context_limit_for_budget(provider: str, pcfg: dict[str, Any]) -> int:
     if configured:
         return configured
     if provider == "nvidia-hosted":
-        return 32768
+        return nvidia_hosted_context_default(str(pcfg.get("current_model") or ""))
     return 65536
 
 
@@ -6124,14 +6142,72 @@ def apply_llm_preset_to_provider(provider: str, pcfg: dict[str, Any], preset_id:
         }
         for token in tokens_by_preset[preset_id]:
             apply_provider_option(provider, pcfg, token)
-    else:
-        native_default = "false" if provider == "nvidia-hosted" else "true"
-        server_limit = upstream_model_context_limit(provider, pcfg) if provider in ("vllm", "self-hosted-nim") else None
-        tokens_by_preset = {
-            "balanced": [
-                "context_window=32768",
-                "reserve=2048",
-                "max_output_tokens=4096",
+    else:
+        native_default = "false" if provider == "nvidia-hosted" else "true"
+        server_limit = upstream_model_context_limit(provider, pcfg) if provider in ("vllm", "self-hosted-nim") else None
+        if provider == "nvidia-hosted":
+            tokens_by_preset = {
+                "balanced": [
+                    "context_window=65536",
+                    "reserve=4096",
+                    "max_output_tokens=4096",
+                    "timeout=300000",
+                    "temperature=0.3",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+                "coding": [
+                    "context_window=65536",
+                    "reserve=4096",
+                    "max_output_tokens=4096",
+                    "timeout=300000",
+                    "temperature=0.2",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+                "fast": [
+                    "context_window=65536",
+                    "reserve=2048",
+                    "max_output_tokens=2048",
+                    "timeout=300000",
+                    "temperature=0.2",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+                "long-context-65k": [
+                    "context_window=131072",
+                    "reserve=8192",
+                    "max_output_tokens=4096",
+                    "timeout=900000",
+                    "temperature=0.3",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+                "large-output": [
+                    "context_window=262144",
+                    "reserve=8192",
+                    "max_output_tokens=8192",
+                    "timeout=1200000",
+                    "temperature=0.3",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+                "reasoning": [
+                    "context_window=262144",
+                    "reserve=8192",
+                    "max_output_tokens=4096",
+                    "timeout=1800000",
+                    "temperature=0.6",
+                    "unset:top_p",
+                    "unset:top_k",
+                ],
+            }
+        else:
+            tokens_by_preset = {
+            "balanced": [
+                "context_window=32768",
+                "reserve=2048",
+                "max_output_tokens=4096",
                 "timeout=300000",
                 "temperature=0.3",
                 "unset:top_p",
@@ -6185,10 +6261,10 @@ def apply_llm_preset_to_provider(provider: str, pcfg: dict[str, Any], preset_id:
                 "timeout=1800000",
                 "temperature=0.6",
                 "unset:top_p",
-                "unset:top_k",
-                f"native={native_default}",
-            ],
-        }
+                "unset:top_k",
+                f"native={native_default}",
+            ],
+            }
         for token in tokens_by_preset[preset_id]:
             if provider == "nvidia-hosted" and token.startswith("native="):
                 continue
@@ -6206,16 +6282,38 @@ def apply_llm_preset_to_provider(provider: str, pcfg: dict[str, Any], preset_id:
         f"{ui_text('apply_preset', lang)}: {label}",
         f"Provider: {provider}; {ui_text('model_family', lang)}: {model_family_text(family, lang)}",
     ]
-    if provider in ("vllm", "self-hosted-nim"):
-        server_limit = upstream_model_context_limit(provider, pcfg)
-        if server_limit:
-            lines.append(f"Server max_model_len: {server_limit}")
-            if preset_id in ("long-context-65k", "large-output") and server_limit < 65536:
-                lines.append("Long-context preset requires restarting the server with --max-model-len 65536 or higher.")
-                lines.append("Client settings were capped to the server-reported context length.")
-        elif preset_id in ("long-context-65k", "large-output"):
-            lines.append("Could not verify server max_model_len; vLLM/NIM must be started with a matching context limit.")
-    return lines
+    if provider in ("vllm", "nvidia-hosted", "self-hosted-nim"):
+        server_limit = upstream_model_context_limit(provider, pcfg)
+        if server_limit:
+            lines.append(f"Server max_model_len: {server_limit}")
+            if preset_id in ("long-context-65k", "large-output") and server_limit < 65536:
+                lines.append("Long-context preset requires restarting the server with --max-model-len 65536 or higher.")
+                lines.append("Client settings were capped to the server-reported context length.")
+        elif preset_id in ("long-context-65k", "large-output"):
+            lines.append("Could not verify server max_model_len; vLLM/NIM must be started with a matching context limit.")
+    if provider in ("vllm", "nvidia-hosted", "self-hosted-nim"):
+        lines.append(
+            "Applied options: "
+            f"context_window={pcfg.get('context_window', 'default')}, "
+            f"reserve={pcfg.get('context_reserve_tokens', 'default')}, "
+            f"max_output_tokens={pcfg.get('max_output_tokens', 'default')}, "
+            f"timeout={pcfg.get('request_timeout_ms', 'default')}ms"
+        )
+    elif provider in ("ollama", "ollama-cloud"):
+        opts = ollama_extra_options(pcfg)
+        lines.append(
+            "Applied options: "
+            f"num_ctx={ollama_num_ctx_status(pcfg)}, "
+            f"num_predict={opts.get('num_predict', 'default')}, "
+            f"timeout={pcfg.get('request_timeout_ms', 'default')}ms"
+        )
+    elif provider == "anthropic":
+        lines.append(
+            "Applied options: "
+            f"max_output_tokens={pcfg.get('max_output_tokens', 'default')}, "
+            f"timeout={pcfg.get('request_timeout_ms', 'default')}ms"
+        )
+    return lines
 
 
 def apply_llm_preset_config(provider: str, preset_id: str) -> list[str]:
@@ -6413,9 +6511,9 @@ def llm_option_panel_rows(provider: str, pcfg: dict[str, Any], lang: str | None
         add("Rate limit RPM", "rate_limit_rpm", pcfg.get("rate_limit_rpm", 40))
         add("Rate limit status", "rate_limit_status", "on" if bool(pcfg.get("rate_limit_status", True)) else "off")
     else:
-        if provider in ("vllm", "self-hosted-nim"):
-            add("Context window", "context_window", pcfg.get("context_window", "default"))
-            add("Context reserve", "context_reserve_tokens", pcfg.get("context_reserve_tokens", "default"))
+        if provider in ("vllm", "nvidia-hosted", "self-hosted-nim"):
+            add("Context window", "context_window", pcfg.get("context_window", "default"))
+            add("Context reserve", "context_reserve_tokens", pcfg.get("context_reserve_tokens", "default"))
         add("Max output tokens", "max_output_tokens", pcfg.get("max_output_tokens", "default"))
         if provider in ("vllm", "nvidia-hosted", "self-hosted-nim"):
             add("Timeout ms", "request_timeout_ms", pcfg.get("request_timeout_ms", "default"))
@@ -6456,13 +6554,38 @@ def llm_option_prompt_default(provider: str, pcfg: dict[str, Any], key: str) ->
     return "" if value is None else str(value)
 
 
-def set_llm_option_config(provider: str, key: str, raw_value: str) -> list[str]:
+def set_llm_option_config(provider: str, key: str, raw_value: str) -> list[str]:
     cfg = load_config()
     pcfg = cfg["providers"][provider]
-    value = raw_value.strip()
-    if not value:
-        return ["Option unchanged."]
-    clear_words = ("default", "unset", "none", "null")
+    value = raw_value.strip()
+    if not value:
+        return ["Option unchanged."]
+    numeric_keys = {
+        "context_window",
+        "context",
+        "max_model_len",
+        "context_reserve_tokens",
+        "reserve",
+        "max_output_tokens",
+        "max_tokens",
+        "maxtoken",
+        "max_token",
+        "num_ctx_min",
+        "num_ctx_max",
+        "num_predict",
+        "timeout",
+        "timeout_ms",
+        "request_timeout",
+        "request_timeout_ms",
+        "rate_limit",
+        "rate_limit_rpm",
+        "rpm",
+        "top_k",
+    }
+    if key in numeric_keys and value.lower() not in ("default", "unset", "none", "null", "0"):
+        if not re.fullmatch(r"\d+", value):
+            return [f"{key}: enter digits only, or use default/unset to clear."]
+    clear_words = ("default", "unset", "none", "null")
     token = f"unset:{key}" if value.lower() in clear_words else f"{key}={value}"
     if provider in ("ollama", "ollama-cloud"):
         apply_ollama_option(pcfg, token)
@@ -8040,14 +8163,16 @@ def render_prelaunch_screen(
     return False
 
 
-def prompt_menu_value(prompt: str, default: str = "", secret: bool = False) -> str:
-    label = f"{prompt}"
-    if default:
-        label += f" [{default}]"
-    label += ": "
-    if sys.stdout.isatty():
-        sys.stdout.write("\033[?25h")
-        sys.stdout.flush()
+def prompt_menu_value(prompt: str, default: str = "", secret: bool = False, restore_tty: Callable[[], None] | None = None, raw_tty: Callable[[], None] | None = None) -> str:
+    label = f"{prompt}"
+    if default:
+        label += f" [{default}]"
+    label += ": "
+    if restore_tty:
+        restore_tty()
+    if sys.stdout.isatty():
+        sys.stdout.write("\033[?25h")
+        sys.stdout.flush()
     sys.stdout.write("\n" + ansi(label, "1;38;5;208"))
     sys.stdout.flush()
     try:
@@ -8055,12 +8180,14 @@ def prompt_menu_value(prompt: str, default: str = "", secret: bool = False) -> s
             value = getpass.getpass("")
         else:
             value = input()
-    finally:
-        if sys.stdout.isatty():
-            sys.stdout.write("\033[?25l")
-            sys.stdout.flush()
-    value = value.strip()
-    return value or default
+    finally:
+        if sys.stdout.isatty():
+            sys.stdout.write("\033[?25l")
+            sys.stdout.flush()
+        if raw_tty:
+            raw_tty()
+    value = value.strip()
+    return value or default
 
 
 def portable_provider_menu() -> int:
@@ -8092,22 +8219,23 @@ def portable_prelaunch_menu() -> int:
     enable_ansi()
     main_idx = 7 if settings_ready_except_api_key() else 0
     panel: str | None = None
-    panel_idx = 0
-    panel_rows: list[str] = []
-    panel_values: list[str] = []
+    panel_idx = 0
+    panel_rows: list[str] = []
+    panel_values: list[str] = []
+    panel_last_idx: dict[str, int] = {}
     checks = preflight_lines()
     messages: list[str] = []
     first_render = True
 
-    def open_panel(name: str) -> None:
-        nonlocal panel, panel_idx, panel_rows, panel_values, messages, first_render
-        cfg = load_config()
-        provider, pcfg = get_current_provider(cfg)
-        panel = name
-        panel_idx = 0
-        if name == "language":
-            panel_rows, panel_values = language_panel_rows(cfg)
-            panel_idx = panel_values.index(cfg.get("language", "en"))
+    def open_panel(name: str) -> None:
+        nonlocal panel, panel_idx, panel_rows, panel_values, messages, first_render
+        cfg = load_config()
+        provider, pcfg = get_current_provider(cfg)
+        panel = name
+        panel_idx = panel_last_idx.get(name, 0)
+        if name == "language":
+            panel_rows, panel_values = language_panel_rows(cfg)
+            panel_idx = panel_values.index(cfg.get("language", "en"))
         elif name == "provider":
             panel_rows, panel_values = provider_panel_rows(cfg)
             panel_idx = panel_values.index(provider)
@@ -8129,15 +8257,19 @@ def portable_prelaunch_menu() -> int:
                 panel_rows, panel_values = [f"Advisor model list failed: {type(exc).__name__}: {exc}", "+ Custom advisor model id..."], []
         elif name == "test":
             panel_rows, panel_values = ["Run compatibility test", "Back"], ["run", "back"]
-        elif name == "options":
-            panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
-        elif name == "preset":
-            panel_rows, panel_values = llm_preset_panel_rows(provider, pcfg, cfg.get("language", "en"))
-
-    def close_panel(next_idx: int | None = None) -> None:
-        nonlocal panel, panel_idx, panel_rows, panel_values, main_idx
-        panel = None
-        panel_idx = 0
+        elif name == "options":
+            panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
+        elif name == "preset":
+            panel_rows, panel_values = llm_preset_panel_rows(provider, pcfg, cfg.get("language", "en"))
+        if panel_rows:
+            panel_idx = max(0, min(panel_idx, len(panel_rows) - 1))
+
+    def close_panel(next_idx: int | None = None) -> None:
+        nonlocal panel, panel_idx, panel_rows, panel_values, main_idx
+        if panel:
+            panel_last_idx[panel] = panel_idx
+        panel = None
+        panel_idx = 0
         panel_rows = []
         panel_values = []
         if next_idx is not None:
@@ -8160,20 +8292,43 @@ def portable_prelaunch_menu() -> int:
             termios.tcsetattr(fd, termios.TCSANOW, new)
         except Exception:
             fd = -1
-    if sys.stdout.isatty():
-        sys.stdout.write("\033[?25l")
-        sys.stdout.flush()
-    try:
+    if sys.stdout.isatty():
+        sys.stdout.write("\033[?25l")
+        sys.stdout.flush()
+    def restore_line_mode() -> None:
+        if old_settings is not None and fd >= 0:
+            try:
+                import termios
+                termios.tcsetattr(fd, termios.TCSANOW, old_settings)
+            except Exception:
+                pass
+
+    def restore_raw_mode() -> None:
+        if old_settings is not None and fd >= 0:
+            try:
+                import termios
+                new = termios.tcgetattr(fd)
+                new[3] = new[3] & ~(termios.ECHO | termios.ICANON)
+                new[6][termios.VMIN] = 1
+                new[6][termios.VTIME] = 0
+                termios.tcsetattr(fd, termios.TCSANOW, new)
+            except Exception:
+                pass
+
+    try:
         while True:
             first_render = render_prelaunch_screen(main_idx, panel, panel_idx, panel_rows, checks, messages, first_render)
             key = read_menu_key(fd) if fd >= 0 else read_menu_key()
-            if panel:
-                if key in ("up", "k"):
-                    panel_idx = (panel_idx - 1) % max(1, len(panel_rows))
-                    continue
-                if key in ("down", "j"):
-                    panel_idx = (panel_idx + 1) % max(1, len(panel_rows))
-                    continue
+            if panel:
+                panel_name = panel
+                if key in ("up", "k"):
+                    panel_idx = (panel_idx - 1) % max(1, len(panel_rows))
+                    panel_last_idx[panel_name] = panel_idx
+                    continue
+                if key in ("down", "j"):
+                    panel_idx = (panel_idx + 1) % max(1, len(panel_rows))
+                    panel_last_idx[panel_name] = panel_idx
+                    continue
                 if key in ("esc", "left", "q"):
                     close_panel()
                     continue
@@ -8198,7 +8353,7 @@ def portable_prelaunch_menu() -> int:
                         close_panel()
                         continue
                     if value == "__custom__" or panel_idx >= len(panel_values):
-                        model_value = prompt_menu_value("Model id or alias")
+                        model_value = prompt_menu_value("Model id or alias", restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                     else:
                         model_value = value
                     if model_value:
@@ -8210,7 +8365,7 @@ def portable_prelaunch_menu() -> int:
                         close_panel()
                         continue
                     if value == "__custom__" or panel_idx >= len(panel_values):
-                        advisor_value = prompt_menu_value("Advisor model id", "deepseek-v4-pro")
+                        advisor_value = prompt_menu_value("Advisor model id", "deepseek-v4-pro", restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                     else:
                         advisor_value = value
                     messages = set_advisor_model_config(advisor_value)
@@ -8220,7 +8375,7 @@ def portable_prelaunch_menu() -> int:
                     if value == "back":
                         close_panel()
                     elif value == "input":
-                        key_value = prompt_menu_value(f"API key for {provider}", secret=True)
+                        key_value = prompt_menu_value(f"API key for {provider}", secret=True, restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                         if key_value:
                             messages = store_api_key_config(provider, key_value)
                             refresh_checks()
@@ -8231,7 +8386,7 @@ def portable_prelaunch_menu() -> int:
                             "nvidia-hosted": "NVIDIA_API_KEY",
                             "ollama-cloud": "OLLAMA_API_KEY",
                         }.get(provider, "API_KEY")
-                        env_name = prompt_menu_value("Environment variable name", default_env)
+                        env_name = prompt_menu_value("Environment variable name", default_env, restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                         key_value = os.environ.get(env_name, "").strip()
                         if key_value:
                             messages = store_api_key_config(provider, key_value)
@@ -8244,7 +8399,7 @@ def portable_prelaunch_menu() -> int:
                         if not key_value:
                             messages = ["Clipboard did not contain readable text."]
                         else:
-                            confirm = prompt_menu_value(f"Clipboard contains {mask_secret(key_value)}. Store it? y/N")
+                            confirm = prompt_menu_value(f"Clipboard contains {mask_secret(key_value)}. Store it? y/N", restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                             if confirm.lower().startswith("y"):
                                 messages = store_api_key_config(provider, key_value)
                             else:
@@ -8260,7 +8415,7 @@ def portable_prelaunch_menu() -> int:
                         close_panel(4)
                     elif value == "edit":
                         default = pcfg.get("base_url") or default_base_url(provider)
-                        url = prompt_menu_value(f"Base URL for {provider}", default)
+                        url = prompt_menu_value(f"Base URL for {provider}", default, restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                         if url:
                             messages = set_base_url_config(provider, url)
                             refresh_checks()
@@ -8289,21 +8444,27 @@ def portable_prelaunch_menu() -> int:
                             messages = set_llm_option_config(provider, value, "false" if current else "true")
                         except Exception as exc:
                             messages = [f"Option update failed: {type(exc).__name__}: {exc}"]
-                        refresh_checks()
-                        cfg = load_config()
-                        provider, pcfg = get_current_provider(cfg)
-                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
-                    else:
-                        default = llm_option_prompt_default(provider, pcfg, value)
-                        entered = prompt_menu_value(f"{value} for {provider} (default/unset clears)", default)
+                        refresh_checks()
+                        cfg = load_config()
+                        provider, pcfg = get_current_provider(cfg)
+                        old_idx = panel_idx
+                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
+                        panel_idx = max(0, min(old_idx, len(panel_rows) - 1))
+                        panel_last_idx["options"] = panel_idx
+                    else:
+                        default = llm_option_prompt_default(provider, pcfg, value)
+                        entered = prompt_menu_value(f"{value} for {provider} (default/unset clears)", default, restore_tty=restore_line_mode, raw_tty=restore_raw_mode)
                         try:
                             messages = set_llm_option_config(provider, value, entered)
                         except Exception as exc:
                             messages = [f"Option update failed: {type(exc).__name__}: {exc}"]
-                        refresh_checks()
-                        cfg = load_config()
-                        provider, pcfg = get_current_provider(cfg)
-                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
+                        refresh_checks()
+                        cfg = load_config()
+                        provider, pcfg = get_current_provider(cfg)
+                        old_idx = panel_idx
+                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
+                        panel_idx = max(0, min(old_idx, len(panel_rows) - 1))
+                        panel_last_idx["options"] = panel_idx
                 elif panel == "preset":
                     if value == "back":
                         open_panel("options")
@@ -8315,12 +8476,14 @@ def portable_prelaunch_menu() -> int:
                         except Exception as exc:
                             messages = [f"Preset failed: {type(exc).__name__}: {exc}"]
                         refresh_checks()
-                        cfg = load_config()
-                        provider, pcfg = get_current_provider(cfg)
-                        panel = "options"
-                        panel_idx = 0
-                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
-                continue
+                        cfg = load_config()
+                        provider, pcfg = get_current_provider(cfg)
+                        panel = "options"
+                        panel_idx = panel_last_idx.get("options", 0)
+                        panel_rows, panel_values = llm_option_panel_rows(provider, pcfg, cfg.get("language", "en"))
+                        panel_idx = max(0, min(panel_idx, len(panel_rows) - 1))
+                        panel_last_idx["options"] = panel_idx
+                continue
 
             if key in ("up", "k"):
                 main_idx = (main_idx - 1) % 10

package/docs/README.ja.md

@@ -47,7 +47,7 @@ vLLM、NVIDIA hosted、self-hosted NIM を選択し、通常の Claude Code 引
 
 Credits: One Ciel LLC
 
-現在のバージョン: `0.1.37`
+現在のバージョン: `0.1.39`
 
 ## 作られた理由
 
@@ -351,6 +351,23 @@ Windows/Linux 管理、クリーンアップスクリプト、定期的なセキ
 
 ## 変更履歴
 
+### 0.1.39
+
+- **メニュー入力修正**: テキスト/数字プロンプトの前に terminal line/echo mode を
+  復元し、prelaunch UI で入力した数字が見えるようにしました。
+- **数字検証の安定化**: 数字オプションへ不正な文字を入れてもメニューが
+  クラッシュせず、案内メッセージを表示します。
+- **プリセット表示改善**: preset 適用後に実際の context、reserve、output、
+  timeout 値をメッセージに表示します。
+
+### 0.1.38
+
+- **ユーザー選択の context window を優先**: NVIDIA hosted の 32K safety cap を
+  削除しました。router は LLM options または headless 設定で選ばれた
+  context window を使い、未設定の場合のみモデル別 fallback を使います。
+- **NVIDIA preset 更新**: NVIDIA hosted preset は 65K から開始し、
+  large-output/reasoning workflow では 256K まで使います。
+
 ### 0.1.37
 
 - **Pseudo tool-call recovery**: NVIDIA/OpenAI-compatible stream 経路で

package/docs/README.ko.md

@@ -47,7 +47,7 @@ NVIDIA hosted, self-hosted NIM을 선택하고, Claude Code의 일반 인자는
 
 Credits: One Ciel LLC
 
-현재 버전: `0.1.37`
+현재 버전: `0.1.39`
 
 ## 왜 만들었나
 
@@ -351,6 +351,23 @@ Windows 이벤트 로그 리뷰, 바이러스/랜섬웨어 침입 시도 정리,
 
 ## 변경 이력
 
+### 0.1.39
+
+- **메뉴 입력 수정**: 텍스트/숫자 프롬프트 전에 터미널 line/echo 모드를 복구하여
+  prelaunch UI에서 입력한 숫자가 보이게 했습니다.
+- **숫자 검증 안정화**: 숫자 옵션에 잘못된 문자를 넣어도 메뉴가 크래시되지 않고
+  안내 메시지를 표시합니다.
+- **프리셋 표시 개선**: preset 적용 후 실제 context, reserve, output, timeout 값을
+  메시지에 표시합니다.
+
+### 0.1.38
+
+- **사용자 선택 context window 우선**: NVIDIA hosted 32K safety cap을 제거했습니다.
+  router는 LLM 옵션 또는 headless 설정에서 선택한 context window를 사용하고,
+  값이 없을 때만 모델별 fallback을 사용합니다.
+- **NVIDIA preset 업데이트**: NVIDIA hosted preset은 65K부터 시작하고,
+  large-output/reasoning 워크플로에서는 256K까지 사용합니다.
+
 ### 0.1.37
 
 - **Pseudo tool-call recovery**: NVIDIA/OpenAI-compatible stream 경로에서

package/docs/README.zh.md

@@ -47,7 +47,7 @@ NIM，并把普通 Claude Code 参数原样传递。
 
 Credits: One Ciel LLC
 
-当前版本: `0.1.37`
+当前版本: `0.1.39`
 
 ## 为什么存在
 
@@ -337,6 +337,22 @@ Hermes 格式模型或部分较旧的 Qwen tool template。
 
 ## 更新日志
 
+### 0.1.39
+
+- **菜单输入修复**：在文本/数字提示前恢复 terminal line/echo mode，
+  prelaunch UI 中输入的数字现在可见。
+- **数字校验更稳**：数字选项输入非法字符时不再让菜单崩溃，而是显示提示消息。
+- **Preset 可见性**：应用 preset 后会显示实际 context、reserve、output、
+  timeout 值。
+
+### 0.1.38
+
+- **优先使用用户选择的 context window**：移除 NVIDIA hosted 的 32K safety cap。
+  router 会使用 LLM options 或 headless 配置中选择的 context window，
+  只有未配置时才使用按模型推断的 fallback。
+- **NVIDIA preset 更新**：NVIDIA hosted preset 从 65K 起步，
+  large-output/reasoning 工作流最高使用 256K。
+
 ### 0.1.37
 
 - **Pseudo tool-call recovery**：NVIDIA/OpenAI-compatible stream 路径现在会

package/docs/manual.md

@@ -10,7 +10,7 @@ Code starts, while passing normal Claude Code arguments through unchanged.
 
 Credits: One Ciel LLC
 
-Current version: `0.1.37`
+Current version: `0.1.39`
 
 ## Install
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@oneciel-ai/claude-any",
-  "version": "0.1.37",
+  "version": "0.1.39",
   "description": "Claude Code provider selector for Anthropic, Ollama, Ollama Cloud, vLLM, NVIDIA hosted, and self-hosted NIM.",
   "license": "MIT",
   "author": "One Ciel LLC",