@nanogpt/private-mode 0.2.11 → 0.2.14

package/README.md

@@ -46,7 +46,7 @@ NanoGPT can see account identity, selected private model, selected TEE target me
 
 NanoGPT's web app can also use these models without running this local proxy. Select an eligible Private Mode model and use the Private Mode control in the model picker. This package is for API clients, CLIs, agents, and other OpenAI-compatible tools.
 
-The web-app toggle is narrower than the proxy path in v1. It supports text chat and model settings, and disables attachments, web search, URL-scraped content, project tools, multi-model chat, Context Memory injection, quick replies, and automatic title generation for private turns.
+The web app supports local image attachments for Private Mode vision models (Kimi K3 and Gemma 4 31B). Image bytes are converted to base64 in the browser and included inside the EHBP-encrypted request; they are not uploaded to NanoGPT object storage. Other attachment types, web search, URL-scraped content, project tools, multi-model chat, Context Memory injection, quick replies, and automatic title generation remain disabled for private turns.
 
 In the hosted web app, decrypted Private Mode turns remain in local browser history. Cloud conversation sync is blocked for Private Mode chats unless password-based end-to-end sync is enabled; the recoverable default sync mode is not used for those chats.
 
@@ -74,10 +74,29 @@ Supported private model IDs include:
 - `private/llama3-3-70b`
 - `private/glm-5-2`
 - `private/glm-5-2:thinking`
+- `private/glm-5-3`
+- `private/glm-5-3-flash`
 - `private/gemma4-31b`
 - `private/gemma4-31b:thinking`
 
-Both GLM 5.2 variants are deployed with a 393,216-token total context limit, with prompt and output sharing that window. Kimi K3 Private has a 256,000-token context limit. DeepSeek V4 Flash 0731 Private has a 1,048,576-token total context limit. The model-list extension `context_length` is the combined prompt-and-output window, while `max_output_tokens` is the output ceiling within that same window. These values are not additive.
+Both GLM 5.2 variants are deployed with a 393,216-token total context limit, with prompt and output sharing that window. GLM 5.3 Private and GLM 5.3 Flash Private each have a 1,048,576-token total context limit and a 131,072-token output ceiling. Kimi K3 Private has a 256,000-token context limit. DeepSeek V4 Flash 0731 Private has a 1,048,576-token total context limit. The model-list extension `context_length` is the combined prompt-and-output window, while `max_output_tokens` is the output ceiling within that same window. These values are not additive.
+
+### Thinking without leaving Private Mode
+
+Keep the same private model ID when enabling thinking. For example:
+
+```json
+{
+  "model": "private/deepseek-v4-flash",
+  "messages": [{ "role": "user", "content": "What is 17 times 19?" }],
+  "reasoning_effort": "high",
+  "stream": true
+}
+```
+
+DeepSeek V4 Flash uses the 0731 release and supports optional thinking. GLM 5.2 and Gemma 4 also support optional thinking. Use `thinking: true` or `enable_thinking: true` to enable it; `thinking: false` disables it for these optional-thinking models. Kimi K3, GLM 5.3, and GLM 5.3 Flash always generate reasoning. GLM 5.3 models accept `low`, `high`, or `max` effort. `reasoning: { "exclude": true }` controls response visibility and does not disable thinking.
+
+Non-streaming reasoning requests may take more than five minutes. The CLI allows the encrypted request to wait for the hosted route's response, with a 30-minute transport limit. The hosted route stops non-streaming generation after at most 29 minutes to leave time before its platform deadline. Use `stream: true` for earlier output, and configure your API client's timeout to allow long generations. Upgrade the local proxy as well as the hosted service to receive both timeout fixes.
 
 Browser requests are locked down by default. The proxy only accepts same-machine clients and rejects browser `Origin` headers that are not explicitly allowed, so a random website or LAN client cannot spend the local `NANOGPT_API_KEY` while the proxy is running. If a local browser app needs to call the proxy directly, allow that exact origin:
 

package/bin/nanogpt-private-mode.js

@@ -4,6 +4,7 @@ import { parseArgs } from 'node:util';
 import { PRIVATE_MODE_PROXY_VERSION } from '../lib/packageInfo.js';
 import { startPrivateModeProxy } from '../lib/server.js';
 import { verifyPrivateModeReceiptCli } from '../lib/verifyReceipt.js';
+import { installPrivateModeTransportTimeouts } from '../lib/transportTimeouts.js';
 
 const { values, positionals } = parseArgs({
   options: {
@@ -69,6 +70,7 @@ if (!apiKey) {
   process.exit(1);
 }
 
+installPrivateModeTransportTimeouts(String(values['api-base']));
 startPrivateModeProxy({
   apiBase: String(values['api-base']),
   apiKey,

package/lib/requestTransforms.js

@@ -29,6 +29,7 @@ const KIMI_K3_TEXT_ENCODER = new TextEncoder();
 const DEEPSEEK_V4_CONTEXT_WINDOW_TOKENS = 1_048_576;
 const DEEPSEEK_V4_MAX_COMPLETION_TOKENS = 1_048_576;
 const DEEPSEEK_V4_CONTEXT_SAFETY_MARGIN_TOKENS = 10;
+const GLM_53_MAX_COMPLETION_TOKENS = 131_072;
 const PRIVATE_TINFOIL_CHAT_COMPLETION_BODY_FIELDS = new Set([
   'chat_template_kwargs',
   'frequency_penalty',
@@ -243,6 +244,14 @@ function normalizeDeepSeekV4ReasoningEffort(value) {
   return 'high';
 }
 
+function normalizeGlm53ReasoningEffort(value) {
+  const normalized = typeof value === 'string' ? value.trim().toLowerCase() : '';
+  if (normalized === 'max' || normalized === 'xhigh') return 'max';
+  if (normalized === 'low' || normalized === 'minimal') return 'low';
+  if (normalized === 'high' || normalized === 'medium') return 'high';
+  return 'max';
+}
+
 function getPrivateModeModelSignals(body, model) {
   return [
     body.model,
@@ -262,8 +271,11 @@ function shouldEnableThinking(body, model) {
   ));
 
   return explicitThinking
+    ?? isThinkingEnabled(body.enable_thinking)
     ?? explicitReasoning
     ?? (body.reasoning_effort !== undefined ? reasoningEffort !== 'none' : undefined)
+    ?? isThinkingEnabled(body.chat_template_kwargs?.thinking)
+    ?? isThinkingEnabled(body.chat_template_kwargs?.enable_thinking)
     ?? hasThinkingSuffix;
 }
 
@@ -419,6 +431,10 @@ function isPrivateModeKimiK3Model(model) {
     });
 }
 
+function isPrivateModeGlm53Model(model) {
+  return model.thinkingMode === 'glm-5.3';
+}
+
 function getPrivateModeFunctionToolName(tool) {
   if (!isPlainObject(tool) || tool.type !== 'function' || !isPlainObject(tool.function)) {
     return undefined;
@@ -513,7 +529,6 @@ export function normalizePrivateModelReasoningControls(body, model) {
 
 export function shouldSuppressPrivateModelReasoning(body, model) {
   const isKimiK3 = isPrivateModeKimiK3Model(model);
-  if (!isKimiK3 && model.thinkingMode !== 'deepseek-v4') return false;
   const reasoningVisibilityOptOut = body.reasoningOptOut === true ||
     body.exposeReasoning === false ||
     body.materializeReasoning === false ||
@@ -528,9 +543,6 @@ export function shouldSuppressPrivateModelReasoning(body, model) {
 
 function resolvePrivateModeKimiK3ReasoningEffort(body) {
   const candidates = [
-    body.reasoningOptOut === true ? false : undefined,
-    body.exposeReasoning === false ? false : undefined,
-    body.materializeReasoning === false ? false : undefined,
     body.enable_thinking,
     body.thinking,
     body.reasoning_effort,
@@ -544,8 +556,7 @@ function resolvePrivateModeKimiK3ReasoningEffort(body) {
   const normalizedEffort = typeof explicitEffort === 'string'
     ? explicitEffort.trim().toLowerCase()
     : undefined;
-  const excludesReasoning = candidates.some((value) => isPlainObject(value) && value.exclude === true);
-  return excludesReasoning ? normalizedEffort ?? 'low' : normalizedEffort ?? 'max';
+  return normalizedEffort ?? 'max';
 }
 
 function applyPrivateModeKimiK3RequestParams(body, model) {
@@ -618,6 +629,20 @@ function clampPrivateModeDeepSeekV4Output(body) {
   );
 }
 
+function clampPrivateModeGlm53Output(body) {
+  const requested = typeof body.max_tokens === 'number' && Number.isFinite(body.max_tokens)
+    ? body.max_tokens
+    : undefined;
+  if (requested === undefined) return;
+  const promptTokens = estimatePrivateModeKimiK3PromptTokens(body);
+  const remainingContext = Math.max(1, DEEPSEEK_V4_CONTEXT_WINDOW_TOKENS - promptTokens - DEEPSEEK_V4_CONTEXT_SAFETY_MARGIN_TOKENS);
+  body.max_tokens = Math.min(
+    requested < 0 ? GLM_53_MAX_COMPLETION_TOKENS : Math.max(1, Math.floor(requested)),
+    GLM_53_MAX_COMPLETION_TOKENS,
+    remainingContext,
+  );
+}
+
 export function createPrivateModeReasoningContentSuppressor() {
   const openTags = ['<think>', '<thinking>', '<previous_reasoning>', '◁think▷'];
   const closeTags = ['</think>', '</thinking>', '</previous_reasoning>', '◁/think▷'];
@@ -1011,10 +1036,12 @@ function normalizeStreamOptions(body) {
 }
 
 export function applyPrivateModelRequestMutations(body, model) {
-  if (isPrivateModeKimiK3Model(model)) {
+  if (isPrivateModeKimiK3Model(model) || isPrivateModeGlm53Model(model)) {
     normalizePrivateModelReasoningControls(body, model);
     normalizePrivateModeKimiK3ReasoningFromMessages(body);
-    normalizePrivateModeKimiK3DynamicToolMessages(body);
+    if (isPrivateModeKimiK3Model(model)) {
+      normalizePrivateModeKimiK3DynamicToolMessages(body);
+    }
   } else {
     stripPrivateModeReasoningFromMessages(body);
   }
@@ -1026,6 +1053,8 @@ export function applyPrivateModelRequestMutations(body, model) {
     applyPrivateModeKimiK3RequestParams(body, model);
   } else if (model.thinkingMode === 'deepseek-v4') {
     clampPrivateModeDeepSeekV4Output(body);
+  } else if (model.thinkingMode === 'glm-5.3') {
+    clampPrivateModeGlm53Output(body);
   }
 
   if (model.thinkingMode === 'gemma') {
@@ -1037,25 +1066,33 @@ export function applyPrivateModelRequestMutations(body, model) {
     delete body.reasoning_effort;
   } else if (
     model.thinkingMode === 'glm-5.2' ||
+    model.thinkingMode === 'glm-5.3' ||
     model.thinkingMode === 'deepseek-v4'
   ) {
-    const thinkingEnabled = shouldEnableThinking(body, model);
+    const thinkingEnabled = model.thinkingMode === 'glm-5.3'
+      ? true
+      : shouldEnableThinking(body, model);
     const requestedReasoningEffort = body.reasoning_effort
-      ?? (isPlainObject(body.reasoning) ? body.reasoning.effort : undefined);
+      ?? (isPlainObject(body.reasoning) ? body.reasoning.effort : undefined)
+      ?? body.chat_template_kwargs?.reasoning_effort;
     body.chat_template_kwargs = {
       ...mergeChatTemplateKwargs(body),
       thinking: thinkingEnabled,
     };
 
-    if (thinkingEnabled) {
-      body.chat_template_kwargs.reasoning_effort =
-        normalizeDeepSeekV4ReasoningEffort(requestedReasoningEffort);
+    if (model.thinkingMode === 'glm-5.3') {
+      // Match the native GLM API and the browser client: low/high/max belongs
+      // at the top level, including on tool turns.
+      body.reasoning_effort = normalizeGlm53ReasoningEffort(requestedReasoningEffort);
+      delete body.chat_template_kwargs.reasoning_effort;
+    } else if (thinkingEnabled) {
+      body.chat_template_kwargs.reasoning_effort = normalizeDeepSeekV4ReasoningEffort(requestedReasoningEffort);
     } else {
       delete body.chat_template_kwargs.reasoning_effort;
     }
 
     delete body.thinking;
-    delete body.reasoning_effort;
+    if (model.thinkingMode !== 'glm-5.3') delete body.reasoning_effort;
   }
 
   stripUnsupportedPrivateTinfoilFields(body);

package/lib/serverErrorNormalization.js

@@ -6,7 +6,7 @@ const EHBP_RESPONSE_NONCE_HEADER = 'ehbp-response-nonce';
 const PRIVATE_MODE_REFUND_NOTICE = 'Any reserved balance will be released or refunded.';
 const MAX_UPSTREAM_ERROR_BODY_BYTES = 16 * 1024;
 const MAX_UPSTREAM_ERROR_MESSAGE_CHARS = 1_000;
-const HIDDEN_PROVIDER_NAME_PATTERN = /\b(?:openrouter|spoke\s*ai|aihubmix|ollama|mimas|comet|azure|digital\s*ocean|axionic|whale\s*ai|langfork)\b/gi;
+const HIDDEN_PROVIDER_NAME_PATTERN = /\b(?:openrouter|spoke\s*ai|aihubmix|mimas|comet|azure|digital\s*ocean|axionic|whale\s*ai|langfork)\b/gi;
 
 export function privateModeProviderFailureMessage(status) {
   if (status === 429) {

package/lib/statusContract.js

@@ -2,6 +2,7 @@ import { createHash } from 'node:crypto';
 
 export const PRIVATE_MODE_FRONTEND_SUPPORTED_FEATURES = Object.freeze([
   'text_chat',
+  'vision_image_inputs',
   'streaming',
   'conversation_history',
   'model_settings',
@@ -9,7 +10,7 @@ export const PRIVATE_MODE_FRONTEND_SUPPORTED_FEATURES = Object.freeze([
 ]);
 
 export const PRIVATE_MODE_FRONTEND_DISABLED_FEATURES = Object.freeze([
-  'attachments',
+  'non_image_attachments',
   'web_search',
   'url_scraped_content',
   'project_chats',

package/lib/transportTimeouts.js

@@ -0,0 +1,28 @@
+import { Agent, setGlobalDispatcher } from 'undici';
+import { PRIVATE_MODE_STREAM_ROUTE_MAX_DURATION_MS } from './constants.js';
+
+export class PrivateModeDispatcher extends Agent {
+  constructor(apiBase) {
+    super();
+    this.chatUrl = new URL(`${apiBase.replace(/\/+$/, '')}/api/v1/private/tinfoil/v1/chat/completions`);
+  }
+
+  dispatch(options, handler) {
+    // EHBP calls native fetch after encryption, without retaining a custom
+    // dispatcher. Scope the CLI's override to its configured encrypted endpoint.
+    // Override dispatch options, not Agent defaults: fetch supplies its own
+    // five-minute timeout values on every dispatch.
+    if (String(options.origin) === this.chatUrl.origin && options.path === this.chatUrl.pathname) {
+      options = {
+        ...options,
+        headersTimeout: PRIVATE_MODE_STREAM_ROUTE_MAX_DURATION_MS,
+        bodyTimeout: PRIVATE_MODE_STREAM_ROUTE_MAX_DURATION_MS,
+      };
+    }
+    return super.dispatch(options, handler);
+  }
+}
+
+export function installPrivateModeTransportTimeouts(apiBase) {
+  setGlobalDispatcher(new PrivateModeDispatcher(apiBase));
+}

package/models/private-tee.json

@@ -76,6 +76,37 @@
     "ownedBy": "nanogpt-private-mode",
     "aliases": ["private/glm-5.2:thinking", "TEE/glm-5-2:thinking", "TEE/glm-5.2:thinking"]
   },
+  {
+    "id": "private/glm-5-3",
+    "name": "GLM 5.3 Private",
+    "upstreamModel": "glm-5-3",
+    "billingModel": "TEE/glm-5.3",
+    "providerPricingModel": "TEE/glm-5.3",
+    "teeTargetModel": "glm-5-3",
+    "maxInputTokens": 1048576,
+    "maxOutputTokens": 131072,
+    "thinkingMode": "glm-5.3",
+    "created": 1788391929,
+    "ownedBy": "nanogpt-private-mode",
+    "aliases": ["private/glm-5.3", "TEE/glm-5-3", "TEE/glm-5.3"]
+  },
+  {
+    "id": "private/glm-5-3-flash",
+    "name": "GLM 5.3 Flash Private",
+    "upstreamModel": "glm-5-3-flash",
+    "billingModel": "TEE/glm-5.3-flash",
+    "providerPricingModel": "TEE/glm-5.3-flash",
+    "teeTargetModel": "glm-5-3-flash",
+    "teeTargetAliases": [
+      "glm5-3-flash"
+    ],
+    "maxInputTokens": 1048576,
+    "maxOutputTokens": 131072,
+    "thinkingMode": "glm-5.3",
+    "created": 1788307200,
+    "ownedBy": "nanogpt-private-mode",
+    "aliases": ["private/glm-5.3-flash", "TEE/glm-5-3-flash", "TEE/glm-5.3-flash"]
+  },
   {
     "id": "private/gemma4-31b",
     "name": "Gemma 4 31B Private",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@nanogpt/private-mode",
-  "version": "0.2.11",
+  "version": "0.2.14",
   "description": "OpenAI-compatible localhost proxy for NanoGPT Private Mode.",
   "type": "module",
   "publishConfig": {
@@ -23,6 +23,7 @@
     "lib/serverErrorNormalization.js",
     "lib/sseResponsePump.js",
     "lib/statusContract.js",
+    "lib/transportTimeouts.js",
     "lib/verifyReceipt.js",
     "models",
     "README.md"
@@ -32,7 +33,8 @@
   },
   "dependencies": {
     "ai": "6.0.220",
-    "tinfoil": "1.1.12"
+    "tinfoil": "1.1.12",
+    "undici": "6.28.0"
   },
   "engines": {
     "node": ">=22"