kiku-agent 1.0.6 → 1.0.7

package/README.md

@@ -65,7 +65,7 @@ Automatic installation is available for global npm installs. Source checkouts, l
 
 After an update or rollback, restart `kiku-agent` and any application hosting `kiku-mcp`. Update settings and rollback history are stored in the normal Kiku user data directory as `update-state.json`; sessions, memory, `.env`, and YouTube cookies are not modified.
 
-For a local checkout before the package is published, use `npm pack` followed by `npm install -g ./kiku-agent-1.0.6.tgz`.
+For a local checkout before the package is published, use `npm pack` followed by `npm install -g ./kiku-agent-1.0.7.tgz`.
 
 ### ✨ Features
 - **LM Studio Integration**: Connects seamlessly to LM Studio's OpenAI-compatible server at `http://127.0.0.1:1234/v1` (default model `google/gemma-4-e4b`, with auto-detection of loaded models).
@@ -101,10 +101,11 @@ npm run kiku-agent
 
 ### Launch with Specific Cloud Provider & Model
 ```bash
-npm run kiku-agent -- --provider openai --model gpt-4o-mini
-npm run kiku-agent -- --provider gemini --model gemini-2.5-flash
-npm run kiku-agent -- --provider groq --model llama-3.1-8b-instant
-npm run kiku-agent -- --provider openrouter --model google/gemini-2.5-flash
+npm run kiku-agent -- --provider openai --model gpt-5.6-luna
+npm run kiku-agent -- --provider gemini --model gemini-3.7-flash
+npm run kiku-agent -- --provider groq --model openai/gpt-oss-120b
+npm run kiku-agent -- --provider deepseek --model deepseek-v4-flash
+npm run kiku-agent -- --provider openrouter --model openrouter/auto
 ```
 
 ### One-shot Question Query
@@ -124,7 +125,7 @@ npm run kiku-agent -- --model google/gemma-4-e4b-Q4_K_M --url http://127.0.0.1:1
 | `-V`, `--version` | Print the installed package version and exit |
 | `-P`, `--provider <name>` | Select active provider (`lmstudio`, `openai`, `gemini`, `groq`, `deepseek`, `openrouter`, `custom`) |
 | `-k`, `--api-key <key>` | Specify API key for cloud provider |
-| `-m`, `--model <id>` | Override model ID (e.g. `gpt-4o-mini`, `google/gemma-4-e4b-Q4_K_M`) |
+| `-m`, `--model <id>` | Override model ID (e.g. `gpt-5.6-luna`, `google/gemma-4-e4b-Q4_K_M`) |
 | `-u`, `--url <url>` | Override provider base URL |
 | `-t`, `--tools` | List all registered tools |
 | `-s`, `--stats` | Print catalog overview statistics |
@@ -227,30 +228,33 @@ Run `/help` inside a session for the full list. Commonly used:
 | Provider | Recommended Active Models | Pricing / Cost Tier | Doc & API Reference | Required ENV Variable |
 |---|---|---|---|---|
 | **LM Studio** (Local) | `google/gemma-4-e4b` *(Q4_K_M, Q6_K, Q8_0)* | **100% Free / Local** (Zero token cost) | [Docs](https://lmstudio.ai/docs) • [API Ref](https://lmstudio.ai/docs/api/local-server) | `LMSTUDIO_BASE_URL` (default: `http://127.0.0.1:1234/v1`) |
-| **OpenAI** | `gpt-4o-mini` *(default)*, `gpt-4o`, `o3-mini` | **Ultra-Cheap** (~$0.15 / 1M input) | [Docs](https://platform.openai.com/docs) • [API Ref](https://platform.openai.com/docs/api-reference) | `OPENAI_API_KEY` |
-| **Google Gemini** | `gemini-2.5-flash` *(default)*, `gemini-2.5-flash-lite`, `gemini-2.5-pro`, `gemini-3.1-pro` | **Ultra-Cheap** ($0.10 / 1M in, 1M-2M context) | [Docs](https://ai.google.dev/docs) • [API Ref](https://ai.google.dev/api) | `GEMINI_API_KEY` |
-| **Groq** | `llama-3.3-70b-versatile` *(default)*, `llama-3.1-8b-instant`, `gpt-oss-120b` | **LPU Ultra-Fast** ($0.05 - $0.59 / 1M in) | [Docs](https://console.groq.com/docs/quickstart) • [API Ref](https://console.groq.com/docs/api-reference) | `GROQ_API_KEY` |
-| **DeepSeek** | `deepseek-chat` *(default, V3)*, `deepseek-reasoner` *(R1 CoT)* | **Ultra-Low Cost** ($0.14 - $0.27 / 1M in) | [Docs](https://api-docs.deepseek.com) • [API Ref](https://api-docs.deepseek.com/api/create-chat-completion) | `DEEPSEEK_API_KEY` |
-| **OpenRouter** | `google/gemini-2.5-flash`, `deepseek/deepseek-chat`, `deepseek/deepseek-r1`, `meta-llama/llama-3.3-70b-instruct`, `qwen/qwen-2.5-72b-instruct`, `:free` | **Flexible Aggregator** (Free & Paid) | [Docs](https://openrouter.ai/docs/quick-start) • [API Ref](https://openrouter.ai/docs/api_reference/overview) | `OPENROUTER_API_KEY` |
+| **OpenAI** | `gpt-5.6-luna` *(default)*, `gpt-5.6-terra`, `gpt-5.6-sol` | **Cost-sensitive → Flagship** ($0.20 - $4.00 / 1M input) | [Docs](https://developers.openai.com/api/docs/models) • [API Ref](https://developers.openai.com/api/reference) | `OPENAI_API_KEY` |
+| **Google Gemini** | `gemini-3.7-flash` *(default)*, `gemini-3.5-flash-lite`, `gemini-3.6-flash` | **Free tier / Stable Flash** ($0.30 - $0.75 / 1M input) | [Docs](https://ai.google.dev/gemini-api/docs/models) • [API Ref](https://ai.google.dev/api/generate-content) | `GEMINI_API_KEY` |
+| **Groq** | `openai/gpt-oss-120b` *(default)*, `openai/gpt-oss-20b` | **LPU Ultra-Fast** ($0.075 - $0.15 / 1M input) | [Docs](https://console.groq.com/docs/quickstart) • [API Ref](https://console.groq.com/docs/api-reference) | `GROQ_API_KEY` |
+| **DeepSeek** | `deepseek-v4-flash` *(default)*, `deepseek-v4-pro` | **V4 / 1M context** ($0.22 - $1.32 / 1M uncached input) | [Docs](https://api-docs.deepseek.com/quick_start/pricing/) • [API Ref](https://api-docs.deepseek.com/api/create-chat-completion) | `DEEPSEEK_API_KEY` |
+| **OpenRouter** | `openrouter/auto` *(default, dynamic catalog)* | **Automatic routing** (billed at routed model rate) | [Docs](https://openrouter.ai/docs/guides/routing/routers/auto-router) • [API Ref](https://openrouter.ai/docs/api/reference/overview) | `OPENROUTER_API_KEY` |
 | **Custom Gateway** | Any OpenAI-compatible server (vLLM, Ollama, TGI, Azure) | **Custom** | [OpenAI Spec](https://platform.openai.com/docs/api-reference) | `LLM_BASE_URL`, `LLM_API_KEY` |
 
 ```bash
 # Launch with a specific cloud provider and model
-npm run kiku-agent -- --provider openai --model gpt-4o-mini
-npm run kiku-agent -- --provider gemini --model gemini-2.5-flash
-npm run kiku-agent -- --provider groq --model llama-3.1-8b-instant
-npm run kiku-agent -- --provider openrouter --model google/gemini-2.5-flash
+npm run kiku-agent -- --provider openai --model gpt-5.6-luna
+npm run kiku-agent -- --provider gemini --model gemini-3.7-flash
+npm run kiku-agent -- --provider groq --model openai/gpt-oss-120b
+npm run kiku-agent -- --provider deepseek --model deepseek-v4-flash
+npm run kiku-agent -- --provider openrouter --model openrouter/auto
 
 # Manage providers interactively within the agent CLI
 /provider list                                     # View all 7 providers and key status
-/provider use groq llama-3.1-8b-instant            # Switch provider and model in one command
+/provider use groq openai/gpt-oss-120b             # Switch provider and model in one command
 /provider models gemini                            # List curated cost-effective models with pricing & context
 /provider info openai                              # Show developer documentation & API details
 /provider test groq                                # Test live endpoint connection and measure latency
 /provider setkey groq gsk-...                      # Set or update API key dynamically at runtime
 /model recommend                                   # View recommended models for active provider
-/model use gpt-4o-mini                             # Switch active model on current provider
-```
+/model use gpt-5.6-luna                            # Switch active model on current provider
+```
+
+At interactive startup, provider switches, and before a one-shot query, the CLI checks the provider's live `/models` catalog. If a configured model has been retired, it selects the first available model from that provider's maintained recommendation list and reports the fallback. Legacy IDs for Groq, DeepSeek, and selected renamed models are normalized automatically.
 
 ---
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "kiku-agent",
-  "version": "1.0.6",
+  "version": "1.0.7",
   "description": "Kiku AI terminal agent and MCP server for Ju Jingyi knowledge, media, and career data.",
   "type": "module",
   "bin": {
@@ -26,10 +26,10 @@
   "kikuUpdate": {
     "channel": "stable",
     "releaseSummary": [
-      "AI research suggestion engine and batch subagent execution",
-      "Enhanced web search fallbacks and missing catalog tool hints",
-      "Updated brand endorsements (Versace) and dual catalog sync",
-      "Secure npm-backed updates and terminal markdown renderer"
+      "Updated production model defaults and live fallback resolution across cloud providers",
+      "Hardened REST API key validation and isolated rate-limit probes",
+      "Fixed discography filtering, MCP configuration isolation, and environment loading order",
+      "Secured YouTube iframe messaging and expanded provider and API regression coverage"
     ],
     "releaseUrl": "https://www.npmjs.com/package/kiku-agent"
   },

package/server/agent/cli.js

@@ -1,10 +1,11 @@
 #!/usr/bin/env node
 
-import readline from 'node:readline';
-import { stdin as input } from 'node:process';
-import { existsSync } from 'node:fs';
-import { spawn } from 'node:child_process';
-import { KikuAgent } from './agentCore.js';
+import readline from 'node:readline';
+import { stdin as input } from 'node:process';
+import { existsSync } from 'node:fs';
+import { spawn } from 'node:child_process';
+import './youtubeAuth.js';
+import { KikuAgent } from './agentCore.js';
 import { LLMClient, LMStudioClient } from './llmClient.js';
 import { toolsSchema, toolHandlers, executeTool } from './tools.js';
 import { config, CLOUD_PROVIDERS } from './config.js';
@@ -1540,15 +1541,23 @@ async function main() {
 
           if (opts.showStats) { printWelcomeHero(); const r = await executeTool('get_system_stats', {}); console.log(JSON.stringify(r, null, 2)); process.exit(0); }
 
-          if (opts.query) {
-            printWelcomeHero();
-            const client = new LLMClient({
+          if (opts.query) {
+            printWelcomeHero();
+            const client = new LLMClient({
               provider: opts.provider || config.activeProvider,
               baseUrl: opts.url,
               apiKey: opts.apiKey,
-              model: opts.model
-            });
-            console.log(`  ${C.bold}${C.gold}QUERY${C.reset} ${C.white}${opts.query}${C.reset}\n`);
+              model: opts.model
+            });
+            const modelManager = new ModelManager(client, { requestedModel: opts.model || client.model });
+            const modelResolution = await modelManager.resolve({ requestedModel: opts.model || client.model });
+            if (!modelResolution.connected || !modelResolution.selectedModel) {
+              throw new Error(modelResolution.error || `No available chat model found for ${client.providerName}`);
+            }
+            if (modelResolution.usedFallback) {
+              console.log(`  ${C.amber}Model ${modelResolution.requestedModel} is unavailable; using ${modelResolution.selectedModel}.${C.reset}\n`);
+            }
+            console.log(`  ${C.bold}${C.gold}QUERY${C.reset} ${C.white}${opts.query}${C.reset}\n`);
             const agent = new KikuAgent({ client, thinkingMode: opts.thinking });
             await executeQueryWithUI(agent, opts.query, agent.session.thinkingMode === 'show', null, 1);
             process.exit(0);
@@ -1665,7 +1674,7 @@ async function main() {
             }
           }
 
-          const modelManager = new ModelManager(client, { requestedModel: opts.model || client.model });
+          const modelManager = new ModelManager(client, { requestedModel: opts.model || client.model });
           const modelResolution = catalogOnlyMode
             ? { connected: false, state: 'catalog_only' }
             : await modelManager.resolve({ requestedModel: opts.model || client.model });
@@ -1685,7 +1694,7 @@ async function main() {
               title: 'Connected & Ready',
               status: 'connected',
               lines: [
-                `• Provider: ${client.providerName} (Local)  • Model: ${modelResolution.selectedModel}`,
+                `• Provider: ${client.providerName}${client.provider === 'lmstudio' ? ' (Local)' : ''}  • Model: ${modelResolution.selectedModel}`,
                 `• Status: Chat & Tool DAG Active  • Tools: ${toolsSchema.length} Loaded`,
               ],
             });

package/server/agent/commandRegistry.js

@@ -123,7 +123,7 @@ export const COMMAND_DETAILS = Object.freeze({
     subcommands: [
       { cmd: 'list', desc: 'List all supported providers, availability, and active status', example: '/provider list' },
       { cmd: 'status', desc: 'Show detailed connection metrics, latency, and credentials for active provider', example: '/provider status' },
-      { cmd: 'use <name> [model]', desc: 'Switch active AI provider and optionally specify initial model', example: '/provider use gemini gemini-2.5-flash' },
+      { cmd: 'use <name> [model]', desc: 'Switch active AI provider and optionally specify initial model', example: '/provider use gemini gemini-3.7-flash' },
       { cmd: 'models [name]', desc: 'Fetch live model catalog from the specified or active provider API', example: '/provider models groq' },
       { cmd: 'setkey <key>', desc: 'Configure and persist API key for the active provider', example: '/provider setkey sk-...' },
       { cmd: 'test', desc: 'Perform a lightweight ping/test call to verify provider responsiveness', example: '/provider test' },
@@ -132,7 +132,7 @@ export const COMMAND_DETAILS = Object.freeze({
     examples: [
       '/provider list',
       '/provider use gemini',
-      '/provider use openai gpt-4o-mini',
+      '/provider use openai gpt-5.6-luna',
       '/provider setkey YOUR_API_KEY',
       '/provider models',
     ],

package/server/agent/commands.js

@@ -1198,7 +1198,7 @@ export const handlers = {
 
       if (!targetProvider) {
         console.log(`\n  ${C.amber}Usage: /provider use <provider_name> [model_id]${C.reset}`);
-        console.log(`  ${C.dim}Example: /provider use groq llama-3.1-8b-instant${C.reset}\n`);
+        console.log(`  ${C.dim}Example: /provider use groq openai/gpt-oss-120b${C.reset}\n`);
         return 'HANDLED';
       }
 
@@ -1206,10 +1206,16 @@ export const handlers = {
         const options = targetModel ? { model: targetModel } : {};
         const info = ctx.client.switchProvider(targetProvider, options);
         if (ctx.agent?.client) ctx.agent.client.switchProvider(targetProvider, options);
-        if (ctx.modelManager) {
-          ctx.modelManager.client = ctx.client;
-          ctx.modelManager.requestedModel = info.model;
-        }
+        if (ctx.modelManager) {
+          ctx.modelManager.client = ctx.client;
+          ctx.modelManager.requestedModel = info.model;
+          const resolution = await ctx.modelManager.resolve({ requestedModel: info.model });
+          if (resolution.connected && resolution.usedFallback) {
+            console.log(`  ${C.amber}• Requested model unavailable; using ${resolution.selectedModel}${C.reset}`);
+          } else if (!resolution.connected) {
+            console.log(`  ${C.amber}• Provider is not ready: ${resolution.error}${C.reset}`);
+          }
+        }
         ctx.baseUrl = info.baseUrl;
 
         console.log(`\n  ${C.emerald}[OK] Switched provider to ${C.bold}${info.providerName}${C.reset} (${info.provider})`);

package/server/agent/config.js

@@ -1,4 +1,31 @@
-export const CLOUD_PROVIDERS = {
+const PROVIDER_MODEL_ALIASES = {
+  openai: {
+    'o1-mini': 'gpt-5.6-luna',
+    'o3-mini': 'gpt-5.6-luna',
+  },
+  openrouter: {
+    auto: 'openrouter/auto',
+  },
+  groq: {
+    'gpt-oss-120b': 'openai/gpt-oss-120b',
+    'gpt-oss-20b': 'openai/gpt-oss-20b',
+  },
+  deepseek: {
+    'deepseek-chat': 'deepseek-v4-flash',
+    'deepseek-reasoner': 'deepseek-v4-flash',
+  },
+  gemini: {
+    'gemini-3.1-pro': 'gemini-3.1-pro-preview',
+  },
+};
+
+export function normalizeProviderModel(providerId, modelId) {
+  const normalizedProvider = String(providerId || '').toLowerCase().trim();
+  const normalizedModel = String(modelId || '').trim();
+  return PROVIDER_MODEL_ALIASES[normalizedProvider]?.[normalizedModel] || normalizedModel;
+}
+
+export const CLOUD_PROVIDERS = {
   lmstudio: {
     id: 'lmstudio',
     name: 'LM Studio (Local)',
@@ -23,18 +50,17 @@ export const CLOUD_PROVIDERS = {
     name: 'OpenAI Cloud',
     baseUrl: process.env.OPENAI_BASE_URL || 'https://api.openai.com/v1',
     apiKey: process.env.OPENAI_API_KEY || '',
-    defaultModel: process.env.OPENAI_MODEL || 'gpt-4o-mini',
+    defaultModel: normalizeProviderModel('openai', process.env.OPENAI_MODEL || 'gpt-5.6-luna'),
     isLocal: false,
     envKey: 'OPENAI_API_KEY',
-    docUrl: 'https://platform.openai.com/docs',
-    apiRefUrl: 'https://platform.openai.com/docs/api-reference',
-    pricingSummary: 'Pay-per-token (gpt-4o-mini: ~$0.15/1M in, $0.60/1M out)',
+    docUrl: 'https://developers.openai.com/api/docs/models',
+    apiRefUrl: 'https://developers.openai.com/api/reference',
+    pricingSummary: 'GPT-5.6 Luna: $0.20/1M in, $1.20/1M out; Terra/Sol available for higher capability',
     features: ['tools', 'streaming', 'vision', 'reasoning', 'system-prompt'],
     recommendedModels: [
-      { id: 'gpt-4o-mini', name: 'GPT-4o Mini', tier: 'ultra-cheap', cost: '$0.15 / 1M in', context: '128k', hasTools: true, description: 'Recommended default. Fast, intelligent, and highly cost-effective with tool use.' },
-      { id: 'gpt-4o', name: 'GPT-4o', tier: 'flagship', cost: '$2.50 / 1M in', context: '128k', hasTools: true, description: 'OpenAI flagship multimodal model for complex multimodal & reasoning tasks.' },
-      { id: 'o3-mini', name: 'o3 Mini', tier: 'reasoning', cost: '$1.10 / 1M in', context: '200k', hasTools: true, description: 'Fast, cost-efficient reasoning model with function calling capabilities.' },
-      { id: 'o1-mini', name: 'o1 Mini', tier: 'reasoning', cost: '$1.10 / 1M in', context: '128k', hasTools: true, description: 'Reasoning model specialized for complex STEM logic and coding.' },
+      { id: 'gpt-5.6-luna', name: 'GPT-5.6 Luna', tier: 'cost-sensitive', cost: '$0.20 / 1M in', context: '1.05M', hasTools: true, description: 'Recommended default for efficient, high-volume multilingual agent workloads.' },
+      { id: 'gpt-5.6-terra', name: 'GPT-5.6 Terra', tier: 'balanced', cost: '$2.00 / 1M in', context: '1.05M', hasTools: true, description: 'Balances intelligence and cost for production reasoning and tool use.' },
+      { id: 'gpt-5.6-sol', name: 'GPT-5.6 Sol', tier: 'flagship', cost: '$4.00 / 1M in', context: '1.05M', hasTools: true, description: 'Flagship model for complex professional reasoning and coding.' },
     ],
   },
   openrouter: {
@@ -42,62 +68,53 @@ export const CLOUD_PROVIDERS = {
     name: 'OpenRouter Cloud',
     baseUrl: process.env.OPENROUTER_BASE_URL || 'https://openrouter.ai/api/v1',
     apiKey: process.env.OPENROUTER_API_KEY || '',
-    defaultModel: process.env.OPENROUTER_MODEL || 'google/gemini-2.5-flash',
-    extraHeaders: {
-      'HTTP-Referer': 'https://jujingyi-kikuuu.vercel.app/',
-      'X-Title': 'Kiku Agent',
-    },
+    defaultModel: normalizeProviderModel('openrouter', process.env.OPENROUTER_MODEL || 'openrouter/auto'),
+    extraHeaders: {
+      'HTTP-Referer': 'https://jujingyi-kikuuu.vercel.app/',
+      'X-OpenRouter-Title': 'Kiku Agent',
+    },
     isLocal: false,
     envKey: 'OPENROUTER_API_KEY',
-    docUrl: 'https://openrouter.ai/docs/quick-start',
-    apiRefUrl: 'https://openrouter.ai/docs/api_reference/overview',
-    pricingSummary: 'Universal aggregator with free tiers & flexible pay-as-you-go',
+    docUrl: 'https://openrouter.ai/docs/guides/routing/routers/auto-router',
+    apiRefUrl: 'https://openrouter.ai/docs/api/reference/overview',
+    pricingSummary: 'Auto Router selects a current model dynamically; usage is billed at the routed model rate',
     features: ['universal-routing', 'tools', 'streaming', 'free-tier', 'fallback'],
     recommendedModels: [
-      { id: 'google/gemini-2.5-flash', name: 'Gemini 2.5 Flash', tier: 'ultra-cheap', cost: '$0.10 / 1M in', context: '1M', hasTools: true, description: 'Super fast with huge 1M context window and native function calling.' },
-      { id: 'deepseek/deepseek-chat', name: 'DeepSeek Chat (V3)', tier: 'ultra-cheap', cost: '$0.14 / 1M in', context: '64k', hasTools: true, description: 'DeepSeek-V3 671B MoE via OpenRouter at ultra-low price.' },
-      { id: 'deepseek/deepseek-r1', name: 'DeepSeek Reasoner (R1)', tier: 'reasoning', cost: '$0.55 / 1M in', context: '64k', hasTools: false, description: 'Full DeepSeek-R1 reasoning with Chain-of-Thought output.' },
-      { id: 'meta-llama/llama-3.3-70b-instruct', name: 'Llama 3.3 70B Instruct', tier: 'balanced', cost: '$0.40 / 1M in', context: '128k', hasTools: true, description: 'State-of-the-art open-weights flagship model.' },
-      { id: 'openai/gpt-4o-mini', name: 'GPT-4o Mini', tier: 'ultra-cheap', cost: '$0.15 / 1M in', context: '128k', hasTools: true, description: 'OpenAI mini model routed via OpenRouter.' },
-      { id: 'qwen/qwen-2.5-72b-instruct', name: 'Qwen 2.5 72B Instruct', tier: 'balanced', cost: '$0.35 / 1M in', context: '128k', hasTools: true, description: 'Top-tier open-weights multilingual and coding model.' },
-      { id: 'meta-llama/llama-3.1-8b-instruct:free', name: 'Llama 3.1 8B (Free)', tier: 'free', cost: '$0.00', context: '128k', hasTools: true, description: 'Free tier Llama 3.1 8B on OpenRouter.' },
+      { id: 'openrouter/auto', name: 'OpenRouter Auto Router', tier: 'dynamic', cost: 'Routed model rate', context: 'Model-dependent', hasTools: true, description: 'Recommended default. Chooses from the current catalog so a retired third-party model does not break startup.' },
     ],
   },
-  groq: {
+  groq: {
     id: 'groq',
     name: 'Groq Cloud',
     baseUrl: process.env.GROQ_BASE_URL || 'https://api.groq.com/openai/v1',
     apiKey: process.env.GROQ_API_KEY || '',
-    defaultModel: process.env.GROQ_MODEL || 'llama-3.3-70b-versatile',
+    defaultModel: normalizeProviderModel('groq', process.env.GROQ_MODEL || 'openai/gpt-oss-120b'),
     isLocal: false,
     envKey: 'GROQ_API_KEY',
     docUrl: 'https://console.groq.com/docs/quickstart',
     apiRefUrl: 'https://console.groq.com/docs/api-reference',
-    pricingSummary: 'Ultra-fast LPU inference (llama-3.1-8b: $0.05/1M, llama-3.3-70b: $0.59/1M)',
+    pricingSummary: 'Ultra-fast LPU inference (GPT-OSS 20B: $0.075/1M in, GPT-OSS 120B: $0.15/1M in)',
     features: ['lpu-speed', 'tools', 'streaming', 'openai-compatible'],
-    recommendedModels: [
-      { id: 'llama-3.3-70b-versatile', name: 'Llama 3.3 70B Versatile', tier: 'balanced', cost: '$0.59 / 1M in', context: '128k', hasTools: true, description: 'Recommended default. 70B flagship quality with fast LPU speed and tool use.' },
-      { id: 'llama-3.1-8b-instant', name: 'Llama 3.1 8B Instant', tier: 'ultra-cheap', cost: '$0.05 / 1M in', context: '128k', hasTools: true, description: 'Blazing fast (~800 tokens/s) and ultra-low cost for quick inquiries.' },
-      { id: 'gpt-oss-120b', name: 'GPT-OSS 120B', tier: 'balanced', cost: '$0.59 / 1M in', context: '128k', hasTools: true, description: 'Open-weights 120B flagship model running with ultra-fast Groq LPU inference.' },
-      { id: 'mixtral-8x7b-32768', name: 'Mixtral 8x7B', tier: 'cheap', cost: '$0.24 / 1M in', context: '32k', hasTools: true, description: 'Mistral MoE architecture with 32k context.' },
-      { id: 'gemma2-9b-it', name: 'Gemma 2 9B IT', tier: 'cheap', cost: '$0.20 / 1M in', context: '8k', hasTools: true, description: 'Google Gemma 2 on Groq LPU hardware.' },
-    ],
+    recommendedModels: [
+      { id: 'openai/gpt-oss-120b', name: 'GPT-OSS 120B', tier: 'flagship', cost: '$0.15 / 1M in', context: '131k', hasTools: true, description: 'Recommended default. Production-grade 120B reasoning model with tool use on Groq.' },
+      { id: 'openai/gpt-oss-20b', name: 'GPT-OSS 20B', tier: 'ultra-cheap', cost: '$0.075 / 1M in', context: '131k', hasTools: true, description: 'Fast production model for lower-cost, high-volume requests.' },
+    ],
   },
   deepseek: {
     id: 'deepseek',
     name: 'DeepSeek Cloud',
-    baseUrl: process.env.DEEPSEEK_BASE_URL || 'https://api.deepseek.com/v1',
+    baseUrl: process.env.DEEPSEEK_BASE_URL || 'https://api.deepseek.com',
     apiKey: process.env.DEEPSEEK_API_KEY || '',
-    defaultModel: process.env.DEEPSEEK_MODEL || 'deepseek-chat',
+    defaultModel: normalizeProviderModel('deepseek', process.env.DEEPSEEK_MODEL || 'deepseek-v4-flash'),
     isLocal: false,
     envKey: 'DEEPSEEK_API_KEY',
     docUrl: 'https://api-docs.deepseek.com',
     apiRefUrl: 'https://api-docs.deepseek.com/api/create-chat-completion',
-    pricingSummary: 'Ultra-low cost (deepseek-chat: $0.14 cached / $0.27 uncached / 1M)',
-    features: ['deepseek-v3', 'deepseek-r1-reasoning', 'tools', 'streaming', 'context-cache'],
+    pricingSummary: 'V4 Flash: $0.22-$0.44/1M uncached input; V4 Pro: $0.66-$1.32/1M (off-peak/peak)',
+    features: ['deepseek-v4', 'thinking', 'tools', 'streaming', 'context-cache', '1m-context'],
     recommendedModels: [
-      { id: 'deepseek-chat', name: 'DeepSeek Chat (V3)', tier: 'ultra-cheap', cost: '$0.14-$0.27 / 1M in', context: '64k', hasTools: true, description: 'Recommended default. DeepSeek-V3 671B MoE, top-tier reasoning, tool calling, fast & ultra-cheap.' },
-      { id: 'deepseek-reasoner', name: 'DeepSeek Reasoner (R1)', tier: 'reasoning', cost: '$0.55 / 1M in', context: '64k', hasTools: false, description: 'DeepSeek-R1 full Chain-of-Thought reasoning with native thinking output.' },
+      { id: 'deepseek-v4-flash', name: 'DeepSeek V4 Flash', tier: 'balanced', cost: '$0.22-$0.44 / 1M uncached in', context: '1M', hasTools: true, description: 'Recommended default. Current V4 model with tool calls and thinking/non-thinking modes.' },
+      { id: 'deepseek-v4-pro', name: 'DeepSeek V4 Pro', tier: 'flagship', cost: '$0.66-$1.32 / 1M uncached in', context: '1M', hasTools: true, description: 'Higher-capability V4 model with tool calls and thinking mode.' },
     ],
   },
   gemini: {
@@ -105,18 +122,17 @@ export const CLOUD_PROVIDERS = {
     name: 'Google Gemini Cloud',
     baseUrl: process.env.GEMINI_BASE_URL || 'https://generativelanguage.googleapis.com/v1beta/openai/',
     apiKey: process.env.GEMINI_API_KEY || '',
-    defaultModel: process.env.GEMINI_MODEL || 'gemini-2.5-flash',
+    defaultModel: normalizeProviderModel('gemini', process.env.GEMINI_MODEL || 'gemini-3.7-flash'),
     isLocal: false,
     envKey: 'GEMINI_API_KEY',
-    docUrl: 'https://ai.google.dev/docs',
-    apiRefUrl: 'https://ai.google.dev/api',
-    pricingSummary: 'Ultra-cheap (gemini-2.5-flash: $0.10/1M in, $0.40/1M out)',
+    docUrl: 'https://ai.google.dev/gemini-api/docs',
+    apiRefUrl: 'https://ai.google.dev/api/generate-content',
+    pricingSummary: 'Gemini 3.7 Flash: free tier available; paid standard $0.75/1M input through 2026',
     features: ['1m-context', 'multimodal', 'tools', 'streaming', 'thinking'],
     recommendedModels: [
-      { id: 'gemini-2.5-flash', name: 'Gemini 2.5 Flash', tier: 'ultra-cheap', cost: '$0.10 / 1M in', context: '1M', hasTools: true, description: 'Recommended default. Advanced reasoning and thinking capabilities with 1M context.' },
-      { id: 'gemini-2.5-flash-lite', name: 'Gemini 2.5 Flash Lite', tier: 'lowest-cost', cost: '$0.075 / 1M in', context: '1M', hasTools: true, description: 'Cost-optimized 2.5 lightweight model for high volume queries.' },
-      { id: 'gemini-2.5-pro', name: 'Gemini 2.5 Pro', tier: 'flagship', cost: '$1.25 / 1M in', context: '2M', hasTools: true, description: 'Deep reasoning and 2M context for complex multimodal analysis.' },
-      { id: 'gemini-3.1-pro', name: 'Gemini 3.1 Pro', tier: 'flagship', cost: '$1.50 / 1M in', context: '2M', hasTools: true, description: 'Next-gen Gemini 3.1 flagship with state-of-the-art coding and multimodal reasoning.' },
+      { id: 'gemini-3.7-flash', name: 'Gemini 3.7 Flash', tier: 'latest-stable', cost: '$0.75 / 1M in (2026)', context: '1M', hasTools: true, description: 'Recommended default. Latest stable Flash model for agentic workflows and multimodal reasoning.' },
+      { id: 'gemini-3.5-flash-lite', name: 'Gemini 3.5 Flash-Lite', tier: 'lowest-cost', cost: '$0.30 / 1M in', context: '1M', hasTools: true, description: 'Stable, cost-efficient model for high-volume translation and simple processing.' },
+      { id: 'gemini-3.6-flash', name: 'Gemini 3.6 Flash', tier: 'stable', cost: '$0.75 / 1M in (2026)', context: '1M', hasTools: true, description: 'Stable previous-generation Flash fallback with function calling and thinking.' },
     ],
   },
   custom: {

package/server/agent/llmClient.js

@@ -1,9 +1,12 @@
-import { config, CLOUD_PROVIDERS } from './config.js';
+import { config, CLOUD_PROVIDERS, normalizeProviderModel } from './config.js';
 
 export class LLMClient {
   constructor(options = {}) {
-    const providerId = (options.provider || config.activeProvider || 'lmstudio').toLowerCase();
-    const providerDef = CLOUD_PROVIDERS[providerId] || CLOUD_PROVIDERS.lmstudio;
+    const providerId = (options.provider || config.activeProvider || 'lmstudio').toLowerCase();
+    const providerDef = CLOUD_PROVIDERS[providerId];
+    if (!providerDef) {
+      throw new Error(`Unknown provider "${providerId}". Supported: ${Object.keys(CLOUD_PROVIDERS).join(', ')}`);
+    }
 
     this.provider = providerId;
     this.providerName = providerDef.name || providerId;
@@ -11,7 +14,10 @@ export class LLMClient {
     this.apiKey = options.apiKey !== undefined
       ? options.apiKey
       : (providerDef.apiKey !== undefined ? providerDef.apiKey : (providerDef.isLocal ? (config.lmStudio.apiKey || 'lm-studio') : ''));
-    this.model = options.model || providerDef.defaultModel || config.lmStudio.model || 'google/gemma-4-e4b';
+    this.model = normalizeProviderModel(
+      providerId,
+      options.model || providerDef.defaultModel || config.lmStudio.model || 'google/gemma-4-e4b'
+    );
     this.timeout = options.timeout || config.lmStudio.requestTimeoutMs || 180000;
     this.extraHeaders = { ...(providerDef.extraHeaders || {}), ...(options.extraHeaders || {}) };
   }
@@ -27,7 +33,7 @@ export class LLMClient {
     this.providerName = providerDef.name || pId;
     this.baseUrl = (options.baseUrl || providerDef.baseUrl).replace(/\/+$/, '');
     this.apiKey = options.apiKey !== undefined ? options.apiKey : (providerDef.apiKey || '');
-    this.model = options.model || providerDef.defaultModel;
+    this.model = normalizeProviderModel(pId, options.model || providerDef.defaultModel);
     this.extraHeaders = { ...(providerDef.extraHeaders || {}), ...(options.extraHeaders || {}) };
     return this.getProviderInfo();
   }
@@ -40,7 +46,7 @@ export class LLMClient {
   setModel(model) {
     const trimmed = String(model || '').trim();
     if (!trimmed) throw new Error('A model ID is required.');
-    this.model = trimmed;
+    this.model = normalizeProviderModel(this.provider, trimmed);
     return this.model;
   }
 
@@ -187,8 +193,9 @@ export class LLMClient {
     onToken = null,
     onThought = null
   }) {
-    const isReasoningModel = /^(o1|o3|deepseek-reasoner)/i.test(this.model) || this.model.includes('reasoner');
-    const isO1OrO3 = /^(o1|o3)/i.test(this.model);
+    const isLegacyReasoningModel = /^(o1|o3|deepseek-reasoner)/i.test(this.model) || this.model.includes('reasoner');
+    const usesCompletionTokenLimit = /^(o1|o3|gpt-5(?:\.|-|$))/i.test(this.model);
+    const rejectsTemperature = /^(o1|o3|gpt-5(?:\.|-|$))/i.test(this.model);
 
     const buildPayload = (includeTools = true, includeTemp = true) => {
       const p = {
@@ -197,17 +204,17 @@ export class LLMClient {
         stream
       };
 
-      if (!isO1OrO3 && includeTemp && temperature !== undefined && temperature !== null) {
-        p.temperature = temperature;
-      }
-
-      if (isO1OrO3) {
-        p.max_completion_tokens = maxTokens;
-      } else {
-        p.max_tokens = maxTokens;
-      }
-
-      if (includeTools && tools && Array.isArray(tools) && tools.length > 0 && !isReasoningModel) {
+      if (!rejectsTemperature && includeTemp && temperature !== undefined && temperature !== null) {
+        p.temperature = temperature;
+      }
+
+      if (usesCompletionTokenLimit) {
+        p.max_completion_tokens = maxTokens;
+      } else {
+        p.max_tokens = maxTokens;
+      }
+
+      if (includeTools && tools && Array.isArray(tools) && tools.length > 0 && !isLegacyReasoningModel) {
         p.tools = tools;
         p.tool_choice = "auto";
       }

package/server/agent/modelManager.js

@@ -1,11 +1,14 @@
-const NON_CHAT_MODEL = /(embed|embedding|rerank|whisper|tts|clip)/i;
+import { normalizeProviderModel } from './config.js';
+
+const NON_CHAT_MODEL = /(embed|embedding|rerank|whisper|tts|clip)/i;
 
 export class ModelManager {
-  constructor(client, options = {}) {
+  constructor(client, options = {}) {
     if (!client) throw new Error('ModelManager requires an LLM client.');
     this.client = client;
-    this.requestedModel = options.requestedModel || client.model;
-    this.fallbackModel = options.fallbackModel || null;
+    this.requestedModel = options.requestedModel || client.model;
+    this.fallbackModel = options.fallbackModel || null;
+    this.fallbackModels = Array.isArray(options.fallbackModels) ? options.fallbackModels : [];
     this.lastResolution = null;
   }
 
@@ -26,8 +29,8 @@ export class ModelManager {
     };
   }
 
-  select(model, availableModels = []) {
-    const requested = String(model || '').trim();
+  select(model, availableModels = []) {
+    const requested = normalizeProviderModel(this.client.provider, String(model || '').trim());
     if (!requested) throw new Error('A model ID is required.');
     if (availableModels.length > 0 && !availableModels.includes(requested)) {
       throw new Error(`Model is not available: ${requested}`);
@@ -35,7 +38,11 @@ export class ModelManager {
     if (NON_CHAT_MODEL.test(requested)) {
       throw new Error(`Model does not appear to support chat: ${requested}`);
     }
-    this.client.model = requested;
+    if (typeof this.client.setModel === 'function') {
+      this.client.setModel(requested);
+    } else {
+      this.client.model = requested;
+    }
     this.requestedModel = requested;
     return requested;
   }
@@ -50,12 +57,22 @@ export class ModelManager {
       return this._record({ state: 'server_online_no_models', connected: true, models, error: `${this.client.providerName || 'LLM server'} is online but no model is loaded.` });
     }
 
-    const requested = options.requestedModel || this.requestedModel;
-    const exact = models.find(id => id === requested);
-    const configuredFallback = models.find(id => id === (options.fallbackModel || this.fallbackModel));
-    const gemma = models.find(id => /gemma/i.test(id) && !NON_CHAT_MODEL.test(id));
-    const chatFallback = models.find(id => !NON_CHAT_MODEL.test(id));
-    const selected = exact || configuredFallback || gemma || chatFallback;
+    const requested = options.requestedModel || this.requestedModel;
+    const exact = models.find(id => id === requested);
+    const configuredFallback = models.find(id => id === (options.fallbackModel || this.fallbackModel));
+    const providerRecommendations = this.client.getProviderInfo?.().recommendedModels || [];
+    const fallbackCandidates = [
+      ...(Array.isArray(options.fallbackModels) ? options.fallbackModels : this.fallbackModels),
+      ...providerRecommendations.map(model => model.id),
+    ];
+    const recommendedFallback = fallbackCandidates
+      .map(candidate => models.find(id => id === candidate))
+      .find(Boolean);
+    const gemma = this.client.provider === 'lmstudio'
+      ? models.find(id => /gemma/i.test(id) && !NON_CHAT_MODEL.test(id))
+      : null;
+    const chatFallback = models.find(id => !NON_CHAT_MODEL.test(id));
+    const selected = exact || configuredFallback || recommendedFallback || gemma || chatFallback;
     if (!selected) {
       return this._record({ state: 'no_chat_model', connected: true, models, error: 'No chat-capable model was found.' });
     }

package/server/agent/youtubeAuth.js

@@ -8,6 +8,7 @@ const __dirname = path.dirname(fileURLToPath(import.meta.url));
 const ENV_PATH = path.resolve(__dirname, '..', '..', '.env');
 const AUTH_CONFIG_PATH = path.join(KIKU_DATA_DIR, 'yt_auth.json');
 const SECURE_COOKIE_PATH = path.join(KIKU_DATA_DIR, 'cookies.txt');
+loadEnv();
 const YTDLP_CMD = process.env.KIKU_YTDLP || 'yt-dlp';
 
 
@@ -374,5 +375,3 @@ export function getCookieStatus() {
     authConfigPath: AUTH_CONFIG_PATH
   };
 }
-
-loadEnv();