New FREE RooCode + Gemini 2.5 Update 🤯

Key Concepts

• Rode 3.11 update
• Google Gemini 2.5 Pro integration
• API key balance
• Project-level mCP configuration
• Export/Import settings
• AI-powered coding
• Claude 3.7 Sonic comparison
• Deep site
• Amazon Nova Act

Rode 3.11 Update and Gemini Integration

The video discusses the new Rode 3.11 update, a free AI agent for coding, highlighting its improved integration with Google's Gemini 2.5 Pro API.

• Faster Edits and Improved Gemini Support: The update offers faster editing, less waiting time, and improved Gemini support with smart retries and fixed escaping, ultimately saving money.
• Cost Savings: The speaker mentions a developer saving $150/day by switching from Claude to the free Gemini API within Rode.
• Project-Level mCP Config: The update introduces project-level mCP configuration, allowing configuration on a project-by-project basis instead of globally.
• API Key Balance: The update is supposed to show API key balances, but the speaker couldn't find this feature working during the initial test.

Practical Demonstration and Testing

The speaker demonstrates the new features and compares Gemini with Claude for coding tasks.

• Setting up Gemini: The speaker shows how to plug in the Google Gemini API key into Visual Studio Code, select the latest Gemini version (2.5 Pro), and enable automatic switch modes and subtask creation.
• SEO Cost Calculator Example: The speaker uses the prompt "create a simple SEO cost calculator" to test Gemini's performance. The calculator is created and functions well, updating automatically when keyword numbers are changed.
• API Key Balance Issue: The speaker attempts to check the API key balance for Anthropic (Claude), but the balance doesn't load, indicating a potential issue with this feature.
• Export/Import Settings: The speaker demonstrates the export/import settings feature, allowing users to share or switch between different coding settings.
• Project-Level mCP Config Demonstration: The speaker shows how to access and configure mCP settings globally and on a project level.

Gemini vs. Claude Comparison

The speaker conducts a side-by-side comparison of Gemini and Claude for creating a landing page.

• AI-Powered Client Success Tracker: The speaker uses the prompt "develop an AI-powered client success tracker" to test the coding capabilities.
• Performance Issues with Gemini: Using Gemini in two different windows leads to hitting credit limits, requiring switching between Open Router and Google Gemini settings.
• Sleep Conversion Focused Landing Page: The speaker tests Gemini with the prompt "create a sleep conversion focused landing page," which is completed quickly.
• Iterative Improvement: The speaker takes a screenshot of the generated page and asks Gemini to "make it more attention-grabbing," demonstrating iterative design.
• Claude's Failure: Claude fails to create a functional landing page, producing errors and a blank page, while Gemini successfully creates a usable page.
• Cost Analysis: Claude costs 40 cents for a failed attempt, while Gemini is free and produces a better result.
• Conclusion: Gemini is deemed faster, easier, cheaper, and better than Claude for the tested tasks.

Deep Site: A Free AI Coder

The speaker introduces Deep site, a free AI coder available on Hugging Face Spaces, and compares it with paid alternatives like bolt.new.

• SEO Cost Calculator Comparison: Deep site generates a more sophisticated and visually appealing SEO cost calculator compared to bolt.new.
• Self-Playing Snake Game: Deep site creates a functional self-playing snake game with customizable settings, while bolt.new produces a basic version with limited features.
• SEO Agency Website: Deep site generates a comprehensive SEO agency website with a logo and detailed content, outperforming bolt.new and Gemini in terms of design and information.
• Deep Seek Direct Comparison: Using Deep seek directly is deemed "trash" as it only writes the structure and headers without coding the HTML.
• Conclusion: Deep site is considered a powerful and free tool for building apps, games, and websites, offering better results than paid alternatives.

Amazon Nova Act: A New AI Agent

The speaker discusses Amazon Nova Act, a new AI model trained to perform actions within a web browser.

• Benchmark Performance: Amazon Nova Act outperforms Claude 3.7 Sonet and OpenAI in benchmarks for screen spar text following and web icon interaction.
• Use Cases: The speaker presents examples of Nova Act's capabilities, such as ordering coffee makers on Amazon, reordering items, and booking hotels.
• Potential Applications: The speaker suggests that Nova Act could be integrated into Alexa for automated tasks and lead to the development of new apps and Chrome extensions.
• Capture Handling: Nova Act can overcome captures, which is a significant advantage.
• GitHub Repo: The speaker mentions the availability of a GitHub repo for Amazon Nova Act, indicating an open-source approach.
• Conclusion: Amazon Nova Act is a powerful AI agent with the potential to automate various tasks and revolutionize how people interact with the web.

Synthesis/Conclusion

The video highlights the advancements in AI-powered coding and automation. The Rode 3.11 update with Gemini integration offers a cost-effective and efficient coding solution. Deep site emerges as a free and powerful alternative to paid AI coders. Amazon Nova Act represents a significant step forward in AI agents, capable of performing complex tasks within a web browser. The speaker encourages viewers to explore these tools and join the AI Profit Boardroom for further learning and support.