Inspiration
WordPress powers 40% of all websites, but it was built with 2004 technology and has a clunky interface. We asked ourselves: What if WordPress was built in 2024 with Large Language Models (LLMs) baked in from day one? What if it could create more than just websites and blogs, but full-fledged web apps with databases, APIs, and front-end components by simply chatting with a bot?
What it does
LlamaPress is a free, open-source platform for creating and managing websites and web applications. It allows users to create and modify their sites through natural language interactions with LlamaBot, an AI-powered chatbot. LlamaPress can: Generate and redesign webpages Create forms and other interactive elements Build full-stack web applications with databases, APIs, and front-end components Customize designs based on user preferences and branding
How we built it
LlamaPress is built with Ruby on Rails and integrates various AI technologies: LlamaBot: The core chatbot interface powered by large language models like OpenAI's GPT and Anthropic's Claude. Fetch AI Agents: Used to implement HTML + Tailwind webpages and generate app scaffolding (database, Active Record Models, APIs, and front-end components). Grok with Llama 80B: Used for low-latency decisions. Prem: Serves as a knowledge base for LlamaBot and provides tutorials to guide users. Toolhouse: Fetches page information to incorporate into prompts. We also plan to implement LangGraph for a better experience in managing conversation flow and decision-making.
Challenges we ran into
Figuring out how to deploy AI agents effectively
Implementing a robust knowledge base for LlamaBot
Managing the complexities of generating full-stack applications through natural language
Handling HTML that gets cut off due to LLM token limits
Accomplishments that we're proud of
Creating a user-friendly interface that allows non-technical users to build complex web applications Integrating multiple AI technologies to create a powerful, versatile platform Developing an open-source alternative to WordPress that leverages cutting-edge AI capabilities
What we learned
The potential of AI agents in web development and content creation The challenges and opportunities in creating a conversational interface for web development The importance of robust knowledge bases and context management in AI-powered tools
What's next for LlamaPress
Implement LangGraph for improved conversation flow and decision-making
Develop a solution for detecting and handling truncated HTML output
Enhance the knowledge base to cover more use cases and provide better guidance
Improve the deployment process for AI agents
Implement a system for storing and applying branding information across generated pages
Expand the capabilities of LlamaPress to handle more complex web applications and integrations
Built With
- fastapi
- fetchai
- groq
- premai
- python
- ruby-on-rails
- toolhouse
Log in or sign up for Devpost to join the conversation.