videonut 1.2.7 → 1.3.0

package/memory/short_term/asset_manifest.md

@@ -1,64 +1,64 @@
-# Asset Manifest: The Transformer Paradox
-
-**Project:** 2017-TRANS-AI
-**Status:** Pending Download
-**Hunter:** Scavenger
-
----
-
-## Visual Assets
-
-### Scene 1: The Lab
-*   **Asset ID:** V01-PAPER-HEADER
-*   **Description:** Header of the paper "Attention Is All You Need".
-*   **Search Query:** "Attention Is All You Need paper pdf header 2017"
-*   **Type:** Image (Screenshot)
-
-*   **Asset ID:** V02-GOOGLE-OFFICE-2017
-*   **Description:** Google Brain office aesthetic, cluttered desk, code on screen.
-*   **Search Query:** "Google Brain office interior 2017 messy desk code monitor stock"
-*   **Type:** Stock Image/B-Roll
-
-### Scene 2: The Publication
-*   **Asset ID:** V03-TWITTER-SCROLL
-*   **Description:** Fast scrolling social media feed, tech twitter reaction.
-*   **Search Query:** "fast scrolling twitter feed stock video blue light"
-*   **Type:** Stock Video
-
-### Scene 3: The Boardroom (Google)
-*   **Asset ID:** V04-BOARDROOM-TENSION
-*   **Description:** Sterile glass meeting room, executives arguing, corporate silhouette.
-*   **Search Query:** "corporate boardroom argument glass walls cinematic stock video"
-*   **Type:** Stock Video
-
-### Scene 4: The Garage (OpenAI)
-*   **Asset ID:** V05-OPENAI-EARLY-DAYS
-*   **Description:** OpenAI early team working, Ilya Sutskever/Greg Brockman candid, hackathon vibe.
-*   **Search Query:** "OpenAI team photo 2016 2017 Ilya Sutskever Greg Brockman coding"
-*   **Type:** Image
-
-### Scene 5: The Scaling Laws
-*   **Asset ID:** V06-SCALING-GRAPH
-*   **Description:** Graph from "Scaling Laws for Neural Language Models" (2020).
-*   **Search Query:** "OpenAI Scaling Laws for Neural Language Models graph loss vs compute"
-*   **Type:** Image
-
-*   **Asset ID:** V07-RACE-CAR
-*   **Description:** Formula 1 car speeding past a slow bus.
-*   **Search Query:** "Formula 1 race car overtaking bus stock video"
-*   **Type:** Stock Video (Metaphor)
-
-### Scene 6: The Empty Chairs (The Authors)
-*   **Asset ID:** V08-AUTHORS-COLLAGE
-*   **Description:** Photos of the 8 authors: Vaswani, Shazeer, Parmar, Uszkoreit, Jones, Gomez, Kaiser, Polosukhin.
-*   **Search Query:** "Attention Is All You Need authors photo collage"
-*   **Type:** Image
-
-### Scene 7: The Conclusion
-*   **Asset ID:** V09-OPENAI-LOGO-SHIFT
-*   **Description:** OpenAI logo (Green/White) fading to Black/Closed.
-*   **Search Query:** "OpenAI logo animation green to black"
-*   **Type:** Motion Graphic/Image
-
----
-*End of Manifest*
+# Asset Manifest: The Transformer Paradox
+
+**Project:** 2017-TRANS-AI
+**Status:** Pending Download
+**Hunter:** Scavenger
+
+---
+
+## Visual Assets
+
+### Scene 1: The Lab
+*   **Asset ID:** V01-PAPER-HEADER
+*   **Description:** Header of the paper "Attention Is All You Need".
+*   **Search Query:** "Attention Is All You Need paper pdf header 2017"
+*   **Type:** Image (Screenshot)
+
+*   **Asset ID:** V02-GOOGLE-OFFICE-2017
+*   **Description:** Google Brain office aesthetic, cluttered desk, code on screen.
+*   **Search Query:** "Google Brain office interior 2017 messy desk code monitor stock"
+*   **Type:** Stock Image/B-Roll
+
+### Scene 2: The Publication
+*   **Asset ID:** V03-TWITTER-SCROLL
+*   **Description:** Fast scrolling social media feed, tech twitter reaction.
+*   **Search Query:** "fast scrolling twitter feed stock video blue light"
+*   **Type:** Stock Video
+
+### Scene 3: The Boardroom (Google)
+*   **Asset ID:** V04-BOARDROOM-TENSION
+*   **Description:** Sterile glass meeting room, executives arguing, corporate silhouette.
+*   **Search Query:** "corporate boardroom argument glass walls cinematic stock video"
+*   **Type:** Stock Video
+
+### Scene 4: The Garage (OpenAI)
+*   **Asset ID:** V05-OPENAI-EARLY-DAYS
+*   **Description:** OpenAI early team working, Ilya Sutskever/Greg Brockman candid, hackathon vibe.
+*   **Search Query:** "OpenAI team photo 2016 2017 Ilya Sutskever Greg Brockman coding"
+*   **Type:** Image
+
+### Scene 5: The Scaling Laws
+*   **Asset ID:** V06-SCALING-GRAPH
+*   **Description:** Graph from "Scaling Laws for Neural Language Models" (2020).
+*   **Search Query:** "OpenAI Scaling Laws for Neural Language Models graph loss vs compute"
+*   **Type:** Image
+
+*   **Asset ID:** V07-RACE-CAR
+*   **Description:** Formula 1 car speeding past a slow bus.
+*   **Search Query:** "Formula 1 race car overtaking bus stock video"
+*   **Type:** Stock Video (Metaphor)
+
+### Scene 6: The Empty Chairs (The Authors)
+*   **Asset ID:** V08-AUTHORS-COLLAGE
+*   **Description:** Photos of the 8 authors: Vaswani, Shazeer, Parmar, Uszkoreit, Jones, Gomez, Kaiser, Polosukhin.
+*   **Search Query:** "Attention Is All You Need authors photo collage"
+*   **Type:** Image
+
+### Scene 7: The Conclusion
+*   **Asset ID:** V09-OPENAI-LOGO-SHIFT
+*   **Description:** OpenAI logo (Green/White) fading to Black/Closed.
+*   **Search Query:** "OpenAI logo animation green to black"
+*   **Type:** Motion Graphic/Image
+
+---
+*End of Manifest*

package/memory/short_term/investigation_dossier.md

@@ -1,31 +1,31 @@
-# Truth Dossier: The Transformer Paradox
-
-## Case File: 2017-TRANS-AI
-**Subject:** How OpenAI Weaponized Google's "Attention Is All You Need"
-**Investigator:** Sherlock
-**Date:** 2025-12-28
-
-## 1. The Sweep (Context)
-In 2017, Google Brain researchers released the seminal paper "Attention Is All You Need," introducing the **Transformer** architecture. At the time, it was a breakthrough for machine translation, replacing the sequential processing of RNNs/LSTMs with a parallel "self-attention" mechanism.
-*   **The Official Narrative:** Google published this to advance the field of NLP (Natural Language Processing) and improve Google Translate.
-*   **The Reality:** They handed the blueprint for AGI (Artificial General Intelligence) to the world on a silver platter.
-
-## 2. The Pivot (The Turning Point)
-While Google integrated Transformers into Search (BERT) and research models (T5, Meena), they remained cautious about releasing generative agents due to "reputational risk" and "safety alignment."
-**OpenAI saw what Google missed:** The Transformer wasn't just good for translation; it was the *only* architecture that scaled predictably with compute.
-*   **The Shift:** OpenAI abandoned their work on LSTMs (which they had heavily invested in) and went all-in on the Transformer Decoder.
-*   **The Bet:** They hypothesized that simply throwing massive data and compute at this specific architecture (GPT - Generative Pre-trained Transformer) would yield emergent reasoning capabilities.
-
-## 3. The Deep Dive (The Hidden Agenda)
-Why did the underdog (OpenAI) beat the giant (Google) using the giant's own sword?
-*   **The Scaling Laws:** In 2020, OpenAI published "Scaling Laws for Neural Language Models." They cracked the code: Performance depends mostly on scale (parameters, data, compute), not just architectural tweaking. Google knew this but was paralyzed by the Innovator's Dilemma—releasing a chatbot might cannibalize their Search ad revenue.
-*   **The Brain Drain:** Of the 8 authors of "Attention Is All You Need," *none* remain at Google. Many founded competitors (Cohere, Character.AI) or joined the ecosystem OpenAI cultivated. OpenAI capitalized on this fragmentation.
-*   **Closed vs. Open:** OpenAI started as "Open" but realized the Transformer's value lay in the *weights* and the *data recipe*, closing their doors (GPT-3 onwards) just as they perfected Google's invention.
-
-## 4. Conclusion (The Verdict)
-OpenAI didn't "steal" the technology; they **recognized its potential for domination** while Google treated it as an academic curiosity. Google built the engine; OpenAI built the race car and drove it off the lot while Google was still checking the tire pressure.
-
-**Motive Uncovered:** OpenAI's usage wasn't just adoption; it was an aggressive strategy to exploit a scalability loophole that Google was too bureaucratic to commercialize.
-
----
-*End of Dossier*
+# Truth Dossier: The Transformer Paradox
+
+## Case File: 2017-TRANS-AI
+**Subject:** How OpenAI Weaponized Google's "Attention Is All You Need"
+**Investigator:** Sherlock
+**Date:** 2025-12-28
+
+## 1. The Sweep (Context)
+In 2017, Google Brain researchers released the seminal paper "Attention Is All You Need," introducing the **Transformer** architecture. At the time, it was a breakthrough for machine translation, replacing the sequential processing of RNNs/LSTMs with a parallel "self-attention" mechanism.
+*   **The Official Narrative:** Google published this to advance the field of NLP (Natural Language Processing) and improve Google Translate.
+*   **The Reality:** They handed the blueprint for AGI (Artificial General Intelligence) to the world on a silver platter.
+
+## 2. The Pivot (The Turning Point)
+While Google integrated Transformers into Search (BERT) and research models (T5, Meena), they remained cautious about releasing generative agents due to "reputational risk" and "safety alignment."
+**OpenAI saw what Google missed:** The Transformer wasn't just good for translation; it was the *only* architecture that scaled predictably with compute.
+*   **The Shift:** OpenAI abandoned their work on LSTMs (which they had heavily invested in) and went all-in on the Transformer Decoder.
+*   **The Bet:** They hypothesized that simply throwing massive data and compute at this specific architecture (GPT - Generative Pre-trained Transformer) would yield emergent reasoning capabilities.
+
+## 3. The Deep Dive (The Hidden Agenda)
+Why did the underdog (OpenAI) beat the giant (Google) using the giant's own sword?
+*   **The Scaling Laws:** In 2020, OpenAI published "Scaling Laws for Neural Language Models." They cracked the code: Performance depends mostly on scale (parameters, data, compute), not just architectural tweaking. Google knew this but was paralyzed by the Innovator's Dilemma—releasing a chatbot might cannibalize their Search ad revenue.
+*   **The Brain Drain:** Of the 8 authors of "Attention Is All You Need," *none* remain at Google. Many founded competitors (Cohere, Character.AI) or joined the ecosystem OpenAI cultivated. OpenAI capitalized on this fragmentation.
+*   **Closed vs. Open:** OpenAI started as "Open" but realized the Transformer's value lay in the *weights* and the *data recipe*, closing their doors (GPT-3 onwards) just as they perfected Google's invention.
+
+## 4. Conclusion (The Verdict)
+OpenAI didn't "steal" the technology; they **recognized its potential for domination** while Google treated it as an academic curiosity. Google built the engine; OpenAI built the race car and drove it off the lot while Google was still checking the tire pressure.
+
+**Motive Uncovered:** OpenAI's usage wasn't just adoption; it was an aggressive strategy to exploit a scalability loophole that Google was too bureaucratic to commercialize.
+
+---
+*End of Dossier*

package/memory/short_term/master_script.md

@@ -1,51 +1,51 @@
-# Master Script: The Transformer Paradox
-
-**Title:** The Engine and the Race Car
-**Logline:** Google built the engine of the future. OpenAI stole the keys and drove it like they stole it.
-**Director:** Spielberg
-**Based on Dossier:** 2017-TRANS-AI
-
----
-
-## Act 1: The Golden Ticket
-
-**Scene 1: The Lab**
-*   **Visual:** Low angle, slow dolly in on a messy desk at Google Brain, 2017. Stacks of papers. A monitor glowing with code. The words "Attention Is All You Need" are typed onto a title page.
-*   **Audio:** Soft, rhythmic keyboard clicking. A low, ambient hum of servers.
-*   **Narrator (V.O.):** It was the greatest gift in the history of artificial intelligence. And Google... just gave it away.
-
-**Scene 2: The Publication**
-*   **Visual:** Montage of a PDF being uploaded. The internet exploding with activity. Twitter feeds scrolling fast.
-*   **Audio:** A "whoosh" sound of a file sending. The chatter of a thousand voices overlapping.
-*   **Narrator (V.O.):** "Attention Is All You Need." A boring title for a weapon. Google thought they were improving translation. In reality? They had just published the blueprints for the atom bomb.
-
-## Act 2: The Hesitation vs. The Bet
-
-**Scene 3: The Boardroom (Google)**
-*   **Visual:** A sterile, glass-walled meeting room. Executives in suits looking worriedly at a whiteboard marked "Risk" and "Ad Revenue."
-*   **Audio:** Muffled arguing. "Safety." "Cannibalization." The ticking of a clock.
-*   **Narrator (V.O.):** Google had the engine. But they were paralyzed. If they released a chatbot, it would kill their Search ads. They were the king afraid to leave his castle.
-
-**Scene 4: The Garage (OpenAI)**
-*   **Visual:** Cut to a grittier, darker room. OpenAI researchers. Hoodies. Late nights. Red Bull cans. They are looking at the same "Attention" paper, but their eyes are wide.
-*   **Audio:** Fast-paced, electronic synth music starts building. The sound of a race car engine revving up.
-*   **Narrator (V.O.):** Across town, the underdog saw something else. They saw a race car. They realized: "If we just make this bigger... it doesn't just translate. It *thinks*."
-
-## Act 3: The Getaway
-
-**Scene 5: The Scaling Laws**
-*   **Visual:** A graph on a screen shooting upward exponentially. The line turns into a road. A sleek car (labeled GPT) tears down it, leaving a clunky bus (labeled Google) in the dust.
-*   **Audio:** The engine roar screams. Tires screeching.
-*   **Narrator (V.O.):** OpenAI stopped trying to be clever. They just poured gas on the fire. Data. Compute. More. They bet the farm on Google's own invention.
-
-**Scene 6: The Empty Chairs**
-*   **Visual:** Back to the Google desk from Scene 1. It's empty. Dust motes dancing in the light. We pan across 8 nameplates—the authors of the paper. One by one, they fade away.
-*   **Audio:** A melancholic wind.
-*   **Narrator (V.O.):** Eight researchers wrote that paper. Today? Not a single one is left at Google. They're gone.
-
-**Scene 7: The Conclusion**
-*   **Visual:** The OpenAI logo glowing neon in the dark. It shifts from "Open" to "Closed."
-*   **Audio:** A heavy metallic door slamming shut. Silence.
-*   **Narrator (V.O.):** Google built the engine. OpenAI won the race. And the irony? They did it by reading Google's manual.
-
-**[FADE TO BLACK]**
+# Master Script: The Transformer Paradox
+
+**Title:** The Engine and the Race Car
+**Logline:** Google built the engine of the future. OpenAI stole the keys and drove it like they stole it.
+**Director:** Spielberg
+**Based on Dossier:** 2017-TRANS-AI
+
+---
+
+## Act 1: The Golden Ticket
+
+**Scene 1: The Lab**
+*   **Visual:** Low angle, slow dolly in on a messy desk at Google Brain, 2017. Stacks of papers. A monitor glowing with code. The words "Attention Is All You Need" are typed onto a title page.
+*   **Audio:** Soft, rhythmic keyboard clicking. A low, ambient hum of servers.
+*   **Narrator (V.O.):** It was the greatest gift in the history of artificial intelligence. And Google... just gave it away.
+
+**Scene 2: The Publication**
+*   **Visual:** Montage of a PDF being uploaded. The internet exploding with activity. Twitter feeds scrolling fast.
+*   **Audio:** A "whoosh" sound of a file sending. The chatter of a thousand voices overlapping.
+*   **Narrator (V.O.):** "Attention Is All You Need." A boring title for a weapon. Google thought they were improving translation. In reality? They had just published the blueprints for the atom bomb.
+
+## Act 2: The Hesitation vs. The Bet
+
+**Scene 3: The Boardroom (Google)**
+*   **Visual:** A sterile, glass-walled meeting room. Executives in suits looking worriedly at a whiteboard marked "Risk" and "Ad Revenue."
+*   **Audio:** Muffled arguing. "Safety." "Cannibalization." The ticking of a clock.
+*   **Narrator (V.O.):** Google had the engine. But they were paralyzed. If they released a chatbot, it would kill their Search ads. They were the king afraid to leave his castle.
+
+**Scene 4: The Garage (OpenAI)**
+*   **Visual:** Cut to a grittier, darker room. OpenAI researchers. Hoodies. Late nights. Red Bull cans. They are looking at the same "Attention" paper, but their eyes are wide.
+*   **Audio:** Fast-paced, electronic synth music starts building. The sound of a race car engine revving up.
+*   **Narrator (V.O.):** Across town, the underdog saw something else. They saw a race car. They realized: "If we just make this bigger... it doesn't just translate. It *thinks*."
+
+## Act 3: The Getaway
+
+**Scene 5: The Scaling Laws**
+*   **Visual:** A graph on a screen shooting upward exponentially. The line turns into a road. A sleek car (labeled GPT) tears down it, leaving a clunky bus (labeled Google) in the dust.
+*   **Audio:** The engine roar screams. Tires screeching.
+*   **Narrator (V.O.):** OpenAI stopped trying to be clever. They just poured gas on the fire. Data. Compute. More. They bet the farm on Google's own invention.
+
+**Scene 6: The Empty Chairs**
+*   **Visual:** Back to the Google desk from Scene 1. It's empty. Dust motes dancing in the light. We pan across 8 nameplates—the authors of the paper. One by one, they fade away.
+*   **Audio:** A melancholic wind.
+*   **Narrator (V.O.):** Eight researchers wrote that paper. Today? Not a single one is left at Google. They're gone.
+
+**Scene 7: The Conclusion**
+*   **Visual:** The OpenAI logo glowing neon in the dark. It shifts from "Open" to "Closed."
+*   **Audio:** A heavy metallic door slamming shut. Silence.
+*   **Narrator (V.O.):** Google built the engine. OpenAI won the race. And the irony? They did it by reading Google's manual.
+
+**[FADE TO BLACK]**

package/package.json

@@ -1,65 +1,62 @@
-{
-  "name": "videonut",
-  "version": "1.2.7",
-  "description": "AI-powered YouTube documentary production pipeline with 10 specialized agents for research, scripting, and asset management",
-  "keywords": [
-    "youtube",
-    "documentary",
-    "ai-agents",
-    "video-production",
-    "gemini-cli",
-    "qwen",
-    "claude",
-    "content-creation",
-    "scriptwriting",
-    "research",
-    "automation"
-  ],
-  "homepage": "https://github.com/konda-vamshi-krishna/videonut",
-  "bugs": {
-    "url": "https://github.com/konda-vamshi-krishna/videonut/issues"
-  },
-  "license": "MIT",
-  "author": {
-    "name": "Vamshi Krishna",
-    "email": "vamshikrishna131437@gmail.com"
-  },
-  "repository": {
-    "type": "git",
-    "url": "git+https://github.com/konda-vamshi-krishna/videonut.git"
-  },
-  "scripts": {
-    "setup": "node setup.js"
-  },
-  "bin": {
-    "videonut": "./bin/videonut.js"
-  },
-  "files": [
-    "agents/",
-    "tools/downloaders/",
-    "tools/validators/",
-    "tools/logging/",
-    "tools/check_env.py",
-    "workflows/",
-    "docs/",
-    "memory/",
-    "scripts/",
-    ".gemini/",
-    ".qwen/",
-    ".claude/",
-    ".antigravity/",
-    "bin/",
-    "setup.js",
-    "config.yaml",
-    "requirements.txt",
-    "workflow_orchestrator.py",
-    "file_validator.py",
-    "README.md",
-    "USER_GUIDE.md",
-    "CONTRIBUTING.md",
-    "LICENSE"
-  ],
-  "engines": {
-    "node": ">=20.0.0"
-  }
+{
+  "name": "videonut",
+  "version": "1.3.0",
+  "description": "AI-powered YouTube documentary production pipeline with 10 specialized agents for research, scripting, and asset management",
+  "keywords": [
+    "youtube",
+    "documentary",
+    "ai-agents",
+    "video-production",
+    "gemini-cli",
+    "qwen",
+    "claude",
+    "content-creation",
+    "scriptwriting",
+    "research",
+    "automation"
+  ],
+  "homepage": "https://github.com/konda-vamshi-krishna/videonut",
+  "bugs": {
+    "url": "https://github.com/konda-vamshi-krishna/videonut/issues"
+  },
+  "license": "MIT",
+  "author": {
+    "name": "Vamshi Krishna",
+    "email": "vamshikrishna131437@gmail.com"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/konda-vamshi-krishna/videonut.git"
+  },
+  "scripts": {
+    "setup": "node setup.js"
+  },
+  "bin": {
+    "videonut": "./bin/videonut.js"
+  },
+  "files": [
+    "agents/",
+    "tools/downloaders/",
+    "tools/validators/",
+    "tools/logging/",
+    "tools/check_env.py",
+    "workflows/",
+    "docs/",
+    "memory/",
+    "scripts/",
+    ".antigravity/",
+    "bin/",
+    "setup.js",
+    "config.yaml",
+    "requirements.txt",
+    "workflow_orchestrator.py",
+    "file_validator.py",
+    "README.md",
+    "USER_GUIDE.md",
+    "CONTRIBUTING.md",
+    "LICENSE"
+  ],
+  "engines": {
+    "node": ">=20.0.0"
+  }
 }

package/requirements.txt

@@ -1,9 +1,9 @@
-# VideoNut Dependencies
-yt-dlp
-playwright
-requests
-beautifulsoup4
-pypdf
-youtube-transcript-api
-youtube-search-python
+# VideoNut Dependencies
+yt-dlp
+playwright
+requests
+beautifulsoup4
+pypdf
+youtube-transcript-api
+youtube-search-python
 httpx

package/setup.js

@@ -297,46 +297,64 @@ async function main() {
         info('  No AI CLI currently installed');
     }
 
-    // Always offer installation choice (Gemini recommended)
+    // Always offer installation choice
     console.log('\n📦 CLI Installation:');
-    console.log('  1. Install Gemini CLI (recommended - by Google)');
-    console.log('  2. Install Claude CLI (by Anthropic)');
-    console.log('  3. Install Qwen CLI (by Alibaba)');
+    console.log('  1. Install BOTH Gemini + Qwen (⭐ RECOMMENDED)');
+    console.log('     → Gemini: Best for content writing & creativity');
+    console.log('     → Qwen: Best for instruction following & agent tasks');
+    console.log('  2. Install Gemini CLI only (by Google)');
+    console.log('  3. Install Qwen CLI only (by Alibaba)');
     if (hasGemini || hasQwen || hasClaude) {
         console.log('  4. Skip - Use existing CLI\n');
     } else {
         console.log('  4. Skip - I will install manually\n');
     }
 
-    const answer = await ask('Enter choice [1 for Gemini]: ');
+    const answer = await ask('Enter choice [1 for BOTH]: ');
 
     if (answer === '2') {
+        // Install Gemini CLI only
         try {
-            info('Installing Claude CLI globally...');
-            execSync('npm install -g @anthropic-ai/claude-code', { stdio: 'inherit' });
-            success('Claude CLI installed! Run "claude" to start.');
+            info('Installing Gemini CLI globally...');
+            execSync('npm install -g @google/gemini-cli', { stdio: 'inherit' });
+            success('Gemini CLI installed! (Best for content writing)');
         } catch (e) {
-            warning('Could not install CLI.');
-            info('Install manually: npm install -g @anthropic-ai/claude-code');
+            warning('Could not install Gemini CLI.');
+            info('Install manually: npm install -g @google/gemini-cli');
         }
     } else if (answer === '3') {
+        // Install Qwen CLI only
         try {
             info('Installing Qwen CLI globally...');
             execSync('npm install -g @qwen-code/qwen-code', { stdio: 'inherit' });
-            success('Qwen CLI installed! Run "qwen" to start.');
+            success('Qwen CLI installed! (Best for instruction following)');
         } catch (e) {
-            warning('Could not install CLI.');
+            warning('Could not install Qwen CLI.');
             info('Install manually: npm install -g @qwen-code/qwen-code');
         }
     } else if (answer !== '4') {
+        // Install BOTH Gemini + Qwen (default)
+        info('Installing BOTH Gemini CLI and Qwen CLI...\n');
+
         try {
             info('Installing Gemini CLI globally...');
             execSync('npm install -g @google/gemini-cli', { stdio: 'inherit' });
-            success('Gemini CLI installed! Run "gemini" to start.');
+            success('Gemini CLI installed! (Best for content writing)');
         } catch (e) {
-            warning('Could not install CLI.');
-            info('Install manually: npm install -g @google/gemini-cli');
+            warning('Could not install Gemini CLI.');
+        }
+
+        try {
+            info('Installing Qwen CLI globally...');
+            execSync('npm install -g @qwen-code/qwen-code', { stdio: 'inherit' });
+            success('Qwen CLI installed! (Best for instruction following)');
+        } catch (e) {
+            warning('Could not install Qwen CLI.');
         }
+
+        console.log('\n✅ Both CLIs installed!');
+        console.log('   Use "gemini" for creative content');
+        console.log('   Use "qwen" for agent/instruction tasks');
     }
 
     // ═══════════════════════════════════════════════════════════════