Skip to content

Introduction

LLM Toolkit helps you build full-featured AI-powered apps quickly on Bubble.io.

This site covers what you need to know to setup and use the plugin.

(Note: this documentation is still largely incomplete. We're working on it! - Aug 12, 2024)

Click Here for the plugin page.

Features

  • 🚫 NO More Waiting: Experience real-time streaming of responses, reminiscent of the official ChatGPT interface. Say goodbye to timeouts and prolonged waits!

  • 💪 Support for 100+ Models: Including OpenAI, Anthropic/Claude, Llama, Mistral, Google Gemini, and more! Unlock unparalleled versatility and scalability for your applications!

  • 🤖 Assistants API with Streaming: Create and use OpenAI Assistants in real-time by streaming your runs. Also supports all other functions of the API.

  • 🧠 Enhanced Memory: Multi-turn conversations made easy with our intuitive message history. Perfect memory that just works.

  • 🛠 Function Calling & API Execution: Supercharge your apps with custom ChatGPT plugins or integrate third-party data sources from APIs.

  • 👁️‍🗨️ Multi-modal support (eg GPT-4o) Elevate your AI experience by integrating GPT-4o, enabling your applications to understand and interpret visual and auditory data alongside text, broadening the scope of interaction and functionality.

  • ⏳ Token Monitoring: Keep an eye on your token consumption for every request you make.

  • ✂ Automatic Truncation: Effortlessly manage messages without worrying about input length.

  • 🛑 Stream Interruption: Regain control and steer the conversation when ChatGPT goes off-course.

  • 🌐 Web Search Capabilities: Built-in web search enables you to extract relevant information from the web in real-time using Google search, or paste URLs/links directly for sites you are interested in.

  • 📂 File Uploads: Seamlessly extract text from uploaded files (.pdf, .pptx, .docx, .csv, and many more types). It's easier than ever to add your own data sources to chats.

  • 📌 Bubble-native Vector Storage & Search: Super-charge your app with "Retrieval Augmented Generation". Introducing Bubble-native embedding vector storage and search, so that you can search through your data without requiring Pinecone, Weaviate, or another third-party vector database service.

  • 🔍 Enhanced Error Handling: With clearer and simpler mechanisms, we've made it easier to troubleshoot and navigate unexpected events.

  • 📌 Custom Headers & bodies: For advanced users, you have full control over the payload and headers of your request.

  • ⚙ Comprehensive API Parameters: Fine-tune your requests using all available parameters, including frequency penalty, presence penalty, and logit bias.

  • 🏠 Custom endpoints Use Microsoft Azure deployments or other hosting providers.

  • 🏠 Self-Hosting: Host your own backend server for complete control over your data.

Quickstart

The simplest way to add ChatGPT-powered streaming to your Bubble app is to add a "Data Container" element to your page. Note that this quick overview is not recommended for production use cases in which you are providing an OpenAI API key.