diff --git a/content/docs/building-with-settlemint/dev-tools/code-studio/ai-code-assistant.mdx b/content/docs/building-with-settlemint/dev-tools/code-studio/ai-code-assistant.mdx new file mode 100644 index 00000000..6c4fd84f --- /dev/null +++ b/content/docs/building-with-settlemint/dev-tools/code-studio/ai-code-assistant.mdx @@ -0,0 +1,304 @@ +--- +title: AI Code Assistant +description: Guide to using the AI Code Assistant in the Code Studio +--- + +## AI Code Assistant + +RooCode is an AI-powered coding assistant integrated into SettleMint’s Code +Studio, replacing the former “AI Genie”. It enhances Code Studio by introducing +a more versatile and powerful AI engine directly in your development +environment. With RooCode, you can generate and improve code using natural +language, leverage multiple AI models for different tasks, and even integrate +custom or local AI instances to meet your project’s needs. This guide will walk +you through what RooCode is, how to set it up, and how to make the most of its +features. + +### What is RooCode and How Does It Enhance Code Studio? + +RooCode is a next-generation AI assistant that lives in your Code Studio editor. +Think of it as your intelligent pair programmer: you can ask it to write code, +explain code, suggest improvements, or even create new project files – all +through simple prompts. Unlike the previous AI Genie (which was tied to a single +AI model), RooCode is built to be provider-agnostic and highly extensible. This +means it can connect to a range of AI models and services: • Multiple AI +Providers: Out of the box, RooCode supports popular AI providers like OpenAI +(GPT models), Anthropic (Claude), Google Vertex AI, AWS Bedrock, and more. +You’re not limited to one AI engine; you can choose the model that best fits +your task for better results. + +- Advanced Context Awareness: RooCode can handle larger context windows and + smarter context management than before. It “remembers” more of your codebase + and conversation history, which helps it generate responses that consider your + entire project scope. In practice, you’ll notice more coherent help even as + your files grow or you switch between different parts of your project. +- Extensibility via MCP: RooCode supports the Model Context Protocol (MCP) , a + framework that lets the AI assistant use external tools and services, + including the (SettleMint MCP + server)[https://console.settlemint.com/documentation/building-with-settlemint/dev-tools/mcp]. + This is a big enhancement for Code Studio – it means the AI can potentially + perform complex operations like looking up information in a knowledge base, + running test suites, or controlling a web browser for web-related tasks, all + from within the coding session. (By default, you’ll only use these features if + you choose to enable or add them, so the environment stays straightforward + unless you need the extra power.) +- Seamless Code Studio Integration: RooCode is fully embedded in SettleMint’s + Code Studio interface. You can access it through the familiar chat or prompt + interface. You can access it through the familiar chat or prompt panel. It + works alongside your code in real-time – for example, you can highlight a + piece of code and ask RooCode to explain or refactor it, and it will provide + the answer or suggestion in seconds. This tight integration means your + development workflow is smoother and more efficient, with AI help always at + your fingertips. + +In summary, RooCode enhances Code Studio by making the AI assistance more +powerful, flexible, and context-aware. Whether you’re a developer looking for +quick code generation or an enterprise user needing compliance-friendly AI, +RooCode adapts to provide the best experience. + +### Step-by-Step Setup and Configuration + +Getting started with RooCode in Code Studio is straightforward. Here’s how to +set up and configure it for your needs: + +1. Open Code Studio: Log in to the SettleMint Console and open your Code Studio + environment. Ensure you have the latest version of the Code Studio where + RooCode is available (if SettleMint releases updates, make sure your + environment is updated). You should notice references to RooCode or AI + Assistant in the IDE interface. +2. Access RooCode Settings: In Code Studio, locate the RooCode settings panel. + This is accessible via an rocket icon the Code Studio toolbar. Click on that + to open the configuration settings. +3. Choose an AI Provider: In the RooCode settings, you’ll see an option to + select your AI provider or model. RooCode supports many providers; common + options include OpenAI, Anthropic, Google Vertex AI, AWS Bedrock, etc. Decide + which AI service you want to use for generating suggestions. For instance, if + you have an OpenAI API key and want to use GPT-4, select “OpenAI.” If you + prefer Anthropic’s Claude, choose “Anthropic” from the dropdown. (You can + change this later or even set up multiple profiles for different + providers.) 4. Enter API Keys/Credentials: After selecting a provider, you’ll + need to provide the API key or credentials for that service: + + - For cloud providers like OpenAI or Anthropic: Enter your API key in the + provided field. You might also need to specify any additional info (for + example, an OpenAI Organization ID if applicable, or select the model + variant from a list). RooCode’s Anthropic integration, for example, will + have a field for the Anthropic API Key and a dropdown to pick which Claude + model to use. + - If you choose OpenAI Compatible or custom endpoints (for instance, via a + service like OpenRouter or Requesty that aggregates models), input the base + URL or choose the service name, and then provide the corresponding API key. + + - For Azure OpenAI or enterprise-specific endpoints: you’ll typically enter + an endpoint URL and an API key (and possibly a deployment name) as required + by that service. RooCode allows configuring a custom base URL for providers + like Anthropic or OpenAI if needed, which is useful for enterprise proxies + or Azure endpoints. + +4. Configure Model and Settings: Once your key is in place, select the exact + model or version you want to use. For example, choose “GPT-4” or a specific + Claude variant from the model dropdown. You can also adjust any optional + settings here: + + - Context Limit or Mode Settings: Some providers/models allow adjusting the + maximum tokens or response length. RooCode might expose these or just + manage them automatically. (By default, it optimizes context usage for + you.) + - MCP and Tools: If you plan to use advanced features, ensure that MCP + servers are enabled in settings (this might be on by default). There may be + an option like “Enable MCP Tools” or similar. If you don’t need these, you + can leave it as is. (Advanced users can add specific MCP server + configurations later—this is optional and not required for basic usage.) + - Profiles (Optional): RooCode supports multiple configuration profiles. You + might see an option to create or switch “API Profiles.” This is useful if + you want to quickly switch between different providers or keys (say one + profile for OpenAI, another for a local model). For now, using the default + or a single profile is fine. + +5. Save and Test: Save your settings (there might be a “Save” button or it may + apply changes immediately). Now test RooCode to confirm it’s working: + - Look for the RooCode chat panel or command input in Code Studio. It might + be a sidebar or bottom panel where you can type a prompt. + - Try a simple prompt like: “Hello RooCode” or ask it to write a snippet, + e.g., “// Prompt: write a Solidity function to add two numbers”. + - RooCode should respond with a code suggestion or answer. If it prompts for + any permissions (like file access, since RooCode can write to files), + approve it to allow the AI to assist with coding tasks. + - If you get an error (e.g., unauthorized or no response), double-check your + API key and internet connectivity, or see if the provider might have usage + limits. Adjust keys or settings as needed. + - With setup complete, you can now fully leverage RooCode in your development + workflow. Use natural language to ask for code, explanations, or + improvements. For example: + - “Create a unit test for the above function.” – RooCode will generate test + code. + - “I’m getting a validation error in this contract, can you help find the + bug?” – RooCode can analyze your code and point out potential issues. + - “Document this function.” – RooCode will write documentation comments + explaining the code. + - You can interact with it as you code, and it will utilize the configured AI + model to assist you. Feel free to adjust the provider or model as you see + what works best for your project. + +### Key Features and Benefits of RooCode + +RooCode brings a rich set of features to improve your development experience in +Code Studio. Here are some of the highlights: + +- 🎯 Multiple AI Models & Providers: Connect RooCode to various AI backends. + You’re not locked into one AI engine – choose from OpenAI’s GPT series, + Anthropic’s Claude, Google’s PaLM/Gemini (via Vertex AI), or even open-source + models through services like Ollama or LM Studio. This flexibility means you + can leverage the strengths of different models (e.g., one might be better at + Ollama or LM Studio. This flexibility means you can leverage the strengths of + different models (e.g., one might be better at code completion, another at + explaining concepts) as needed. +- 📚 Advanced Context Management: RooCode is designed to handle large codebases + and lengthy conversations more gracefully. It uses intelligent context + management to include relevant parts of your project when generating answers. + For you, this means less time spent copy-pasting code to show the AI – RooCode + will automatically consider the files you’re working on and recent + interactions. The result is more informed suggestions that truly understand + your project’s context. +- 🤖 MCP (Model Context Protocol) Support: One of the standout advanced features + is RooCode’s ability to use MCP. This allows the AI assistant to interface + with external tools and services in a standardized way . For example, with an + appropriate MCP server configured, RooCode could perform a task like searching + your company’s knowledge base, querying a database for a value, or running a + custom script – all triggered by an AI command. This extends what the AI can + do beyond text generation, turning it into a mini agent that can act on your + behalf. (This is an optional power-user feature; you can use Code Studio and + RooCode fully without ever touching MCP, but it’s there for those who need to + integrate with other systems.) +- 🛠 In-Editor Tools & Actions: RooCode comes with a variety of built-in + capabilities accessible directly in the editor. It can read from and write to + files in your project (with your permission), meaning it can create new code + files or modify existing ones when you accept its suggestions. It can execute + terminal commands in the Code Studio environment – useful for running tests or + compiling code to verify solutions. It even has the ability to control a + browser or other tools via MCP, as mentioned. These actions help automate + routine tasks: imagine generating code and then automatically running your + test suite to verify it, all through AI assistance. +- 🔒 Customization & Control: Despite its power, RooCode gives you control over + the AI’s behavior. You can set custom instructions (for example, telling the + AI about project-specific guidelines or coding style preferences). You can + also adjust approval settings – e.g., require manual approval every time + RooCode tries to write to a file or run a command, or relax this for trusted + actions to speed up your workflow. For enterprise scenarios, features like + disabling MCP entirely or restricting certain actions are available for + compliance (administrators can centrally manage these policies). This balance + ensures you get helpful automation without sacrificing oversight. +- 🚀 Continuous Improvement: RooCode is regularly updated with performance + improvements and new features. Being a part of the SettleMint platform means + it’s tested for our specific use cases (like blockchain and smart contract + development) and tuned for reliability. Expect faster responses and new + capabilities over time – for instance, support for the latest AI models as + they become available, improved prompt handling, and more. All these benefits + come to you automatically through platform updates. + +Together, these features make RooCode a robust AI co-developer. You’ll find that +repetitive tasks get easier, complex tasks become more approachable with AI +guidance, and your team’s overall development speed and quality can increase. + +### Integrating Personal API Keys and Enterprise/Local Instances + +One of the great advantages of RooCode is its flexibility in how it connects to +AI models. Depending on your needs, you can either use personal API keys for +public AI services, or leverage local/enterprise instances for more control. +Here’s how to manage those scenarios: + +- Using Your Own API Keys: If you have your own accounts with AI providers (such + as an OpenAI API subscription or access to Anthropic’s Claude), you can plug + those credentials into RooCode. In the RooCode settings profile, select the + provider and enter your API key (as described in the setup steps). This will + make Code Studio use your allotment of that AI service for all AI completions + and chats. The benefit is that you can tailor which model and version you use + (and often get the newest models immediately), and you have full visibility + into your usage via the provider’s dashboard. For instance, you might use your + OpenAI key to get GPT-4’s latest features. RooCode will respect any rate + limits or quotas on your key, and you’ll be billed by the provider according + to your plan with them (if applicable). This approach is ideal for individual + power users or teams who want the best models and are okay managing their own + API costs. + +- Enterprise API Integrations: Enterprises often have special arrangements or + requirements for AI usage – such as using Azure OpenAI Service, deploying + models via AWS Bedrock, or using a private endpoint hosted in a secure + environment. RooCode supports these cases. You can configure a custom base URL + and API key to point RooCode to your enterprise’s AI endpoint. For example, if + your company uses Azure OpenAI, you’d select “OpenAI Compatible” and provide + the Azure endpoint URI and key. Similarly, for AWS Bedrock, choose the Bedrock + option and enter the necessary credentials. By doing so, all AI requests from + Code Studio will route through those enterprise channels, ensuring compliance + with your org’s data policies (no data leaves your approved environment). This + is crucial for sectors with strict data governance – you get the convenience + of AI coding assistance while keeping data management in line with internal + rules. + +- Local Instances (Offline/On-Premises Use): RooCode can also work with local AI + models running on your own hardware. This is a powerful feature if you need + full offline capability or extra privacy. Using a tool like Ollama or LM + Studio via AWS Bedrock, or using a private endpoint hosted in a secure + environment. RooCode supports these cases. You can configure a custom base URL + and API key to point RooCode to your enterprise’s AI endpoint. For example, if + your company uses Azure OpenAI, you’d select “OpenAI Compatible” and provide + the Azure endpoint URI and key. Similarly, for AWS Bedrock, choose the Bedrock + option and enter the necessary credentials. By doing so, all AI requests from + Code Studio will route through those enterprise channels, ensuring compliance + with your org’s data policies (no data leaves your approved environment). This + is crucial for sectors with strict data governance – you get the convenience + of AI coding assistance while keeping data management in line with internal + rules. + +- Local Instances (Offline/On-Premises Use): RooCode can also work with local AI + models running on your own hardware. This is a powerful feature if you need + full offline capability or extra privacy. Using a tool like Ollama or LM + Studio , you can host language models on a local server that mimics the + OpenAI API. In RooCode’s settings, you would choose a “Local” provider option + (for instance, LM Studio appears as an option) and set the base URL to your + local server (often something like http://localhost:PORT with no API key + needed or a token if the local server requires one). Once configured, RooCode + will send all requests to the local model, meaning your code and queries never + leave your machine. Keep in mind, running local models may require a powerful + computer, and the AI’s performance depends on the model you use (some + open-source models are smaller than the big cloud ones). Still, this option is + fantastic for experimentation, working offline, or ensuring absolute + confidentiality for sensitive code. + +- Switching and Managing Configurations: Through RooCode’s configuration + profiles feature , you can maintain multiple setups. For instance, you might + have one profile called “Personal-OpenAI” with your OpenAI key and GPT-4, + another called “Enterprise-Internal” for your company’s endpoint, and a third + called “Local-LLM” for a model on your machine. In Code Studio, you can + quickly switch between these depending on the project or context. This + flexibility means you’re never locked in – you can always choose the best + route for AI assistance on a case-by-case basis. + +> Tip: Always ensure that when using external API keys or services, you follow +> the provider’s usage policies and secure your keys. Never commit API keys into +> your code repositories. Set them via the Code Studio interface or environment +> variables if supported. SettleMint’s platform will store any keys you enter in +> a secure way, but it’s good practice to manage and rotate keys periodically. +> For enterprise setups, work with your system administrators to obtain the +> correct endpoints and credentials. + +By integrating your own keys or instances with RooCode, you essentially bring +your preferred AI brain into SettleMint’s Code Studio. This empowers you to use +the AI on your terms – whether prioritizing cost, performance, or compliance. +It’s all about giving you the choice. + +### Conclusion and Next Steps + +RooCode dramatically expands the AI capabilities of SettleMint Code Studio, +making it a versatile assistant for blockchain development and beyond. We’ve +covered what RooCode is, how to get it up and running, its key features, and how +to tailor it to your environment. As you start using RooCode, you may discover +new ways it can help in your daily coding tasks – don’t hesitate to explore +features like custom modes or ask RooCode itself for tips on how it can assist +you best! + +For more detailed technical information, troubleshooting, and advanced tips, +check out the (official RooCode documentation)[https://docs.roocode.com]. The +RooCode community is also active – you can find resources like FAQ pages or +community forums (e.g., RooCode’s Discord or subreddit) via the documentation +site if you’re interested in deep dives or sharing experiences. diff --git a/content/docs/building-with-settlemint/dev-tools/code-studio/meta.json b/content/docs/building-with-settlemint/dev-tools/code-studio/meta.json new file mode 100644 index 00000000..4744a40f --- /dev/null +++ b/content/docs/building-with-settlemint/dev-tools/code-studio/meta.json @@ -0,0 +1,4 @@ +{ + "label": "Code Studio", + "pages": ["code-studio", "ai-code-assistant"] +} diff --git a/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/ai-plugin.mdx b/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/ai-plugin.mdx deleted file mode 100644 index 9f9040f0..00000000 --- a/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/ai-plugin.mdx +++ /dev/null @@ -1,49 +0,0 @@ ---- -title: Generative AI plugin -description: Generative AI plugin ---- - -The integration of an OpenAI GPT plugin into the Smart Contract Set of the -SettleMint platform brings forth significant advantages for the fast development -of new smart contracts. Generative AI technology empowers developers by -providing them with automated code suggestions, smart contract templates, and -code completion capabilities. This accelerates the development process by -reducing manual effort and improving code efficiency. - -Additionally, the interactive debugging and explanation features offered by the -GPT plugin greatly benefit developers of all skill levels. Beginners can -leverage the plugin's explanations to understand complex concepts, improve their -coding skills, and gain insights into best practices. Experienced developers can -utilize the debugging features to quickly identify and resolve issues, leading -to more robust and error-free smart contracts. - -Overall, the integration of the OpenAI GPT plugin into the SettleMint platform's -Smart Contract IDE combines the power of generative AI and interactive debugging -to enhance productivity, code quality, and learning opportunities for developers -in the blockchain space. - -You can find this plugin in the sidebar (magic lamp icon), in context menu and -the action panel. The first time you use the plugin you will get asked for an -OpenAI API key. Depending on your OpenAI API access you can use the default GPT -3.5 model or the GPT 4 model which is slower, but more accurate. Check out the -settings in the IDE to configure this and other settings (like using your own -Azure GPT service) - -Here are some prompts to get you started: - -- Write a Solidity ERC20 token representing a bond, based on OpenZeppelin. - Include a bond repayment and coupon payments. -- Write a Solidity ERC721 token representing real estate, based on OpenZeppelin - and allow fractional ownership of this NFT in the form of ERC20 tokens. -- Write The Graph indexing code that indexes an ERC721 token on the Ethereum - Mainnet. - - diff --git a/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/meta.json b/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/meta.json index 210028c2..17547fad 100644 --- a/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/meta.json +++ b/content/docs/building-with-settlemint/dev-tools/code-studio/smart-contract-sets/meta.json @@ -4,7 +4,6 @@ "smart-contract-sets", "smart-contract-templates", "add-smart-contract-set", - "deploying-a-contract", - "ai-plugin" + "deploying-a-contract" ] } diff --git a/content/release-notes/agentic-code-assistant.mdx b/content/release-notes/agentic-code-assistant.mdx new file mode 100644 index 00000000..4337928b --- /dev/null +++ b/content/release-notes/agentic-code-assistant.mdx @@ -0,0 +1,92 @@ +--- +title: "Agentic code assistant in the Code Studio" +description: "The agentic code assistant is now available in the Code Studio." +author: "SettleMint" +date: 2025-03-12 +--- + +## Agentic code assistant in the Code Studio + + + +We are excited to announce that the AI Genie in Code Studio has been upgraded to +RooCode, a more powerful AI coding assistant. This upgrade brings broader AI +capabilities to our platform, benefiting all users – from those new to Code +Studio to advanced developers and enterprise teams. RooCode expands your AI +toolkit with access to multiple AI providers (including Anthropic’s models ) +and introduces smarter context management and enterprise-ready integration +options. Here’s what the RooCode launch means for you: + +### A Boost for All Code Studio Users + +RooCode makes Code Studio’s AI assistance more flexible, intelligent, and +user-friendly. Instead of being limited to a single AI model, RooCode lets you +tap into a variety of cutting-edge AI models for code generation and guidance. +In practice, this means faster, more accurate code suggestions and support +across use cases – whether you’re crafting a Solidity smart contract or +debugging an integration. The AI is context-aware, so it can consider more of +your project’s code and history when helping you, leading to responses that feel +more relevant to what you’re working on. The interface remains as easy as ever: +simply ask RooCode for help or let it suggest improvements as you code. No +advanced setup needed to start seeing benefits. In short, every SettleMint user +will notice a smarter, more capable AI assistant ready to accelerate your +blockchain development projects. + +### Empowering Developers with Advanced AI Options + +For developers, RooCode unlocks unprecedented control and choice in your +AI-assisted workflow. You can now seamlessly integrate multiple AI providers and +models into Code Studio. Want to use OpenAI’s latest GPT model for one task and +switch to Anthropic’s Claude for another? RooCode makes that possible with a few +clicks, all within the same environment. It supports Model Context Protocol +(MCP) servers, meaning you can harness enhanced AI processing capabilities and +even enable the AI to use external tools or data sources as part of coding +tasks. This translates into smarter code completion, automated documentation, +and more robust debugging assistance. RooCode’s advanced context management +ensures the AI remembers and utilizes larger portions of your codebase, so it +can handle complex, multi-file projects and long-running sessions without losing +track. Developers can also customize their AI setup: define profiles for +different projects, tweak the AI’s behavior with custom instructions, or even +choose local AI models for offline use. In short, RooCode isn’t just an upgrade +– it’s a toolkit for power users to supercharge development productivity with +AI. + +### Enterprise-Grade Integration and Control + +Enterprise clients will especially appreciate that RooCode was built with +security, configurability, and scalability in mind. You have the option to +integrate your own API keys for AI providers or use dedicated enterprise AI +instances, ensuring compliance with internal policies and full control over how +and where your code data is processed. For example, if your organization uses a +private cloud OpenAI endpoint or Anthropic’s enterprise API, Code Studio can be +configured to route requests through those credentials seamlessly. RooCode’s +support for local deployments means you can even run AI models on-premises (or +within your VPC) via supported frameworks like Ollama or LM Studio  , keeping +sensitive code completely in-house. The introduction of MCP support allows +enterprises to extend RooCode with custom tools – connecting the AI assistant to +internal knowledge bases, databases, or devops pipelines – all under your +control. Advanced context management not only improves the developer experience +but also ensures that the AI can function within the boundaries you set (for +instance, respecting project-specific data privacy rules or utilizing extended +context windows for large enterprise codebases). With RooCode, enterprises get a +future-proof AI assistant that adapts to corporate needs: integrate it into your +single sign-on and governance workflows, monitor usage with full transparency, +and scale AI assistance across development teams without compromising on +security or performance. + +Ready to experience RooCode’s capabilities firsthand? It’s live now in your +SettleMint Code Studio – simply log in and start coding to try it out. We +encourage everyone to give RooCode a spin at the SettleMint Console and explore +the rich documentation and guides available on +[the RooCode docs](http://docs.roocode.com) for deeper technical details. This +upgrade is a big step in our mission to empower you with the best tools for +blockchain development. We can’t wait to see what you build with the Code Studio +as your smart contract development partner. Happy coding! 👩‍💻👨‍💻