RAG Chatbot: AI that answers using your published content

RAG Chatbot is a new AI-powered extension that lets you add an intelligent assistant to your app. Unlike traditional chatbots that rely on generic responses, RAG Chatbot uses your own published content to deliver accurate and relevant answers to your users' questions.

RAG, which stands for Retrieval-Augmented Generation, is an AI technique that allows a chatbot to answer questions by relying directly on existing content, rather than on a general-purpose memory. In the case of RAG Chatbot, this content comes from what you’ve already published in your app.

The process starts by converting your content into "embeddings"—mathematical vector representations that capture meaning and context. These embeddings are generated using a language model trained to understand natural language. When a user asks a question, the same model is used to convert the question into an embedding as well. This allows both the content and the question to be expressed in the same language: a shared mathematical space where semantic similarities can be measured.

The system then compares the user’s question with your indexed content and identifies the most relevant sections. At that point, a Large Language Model (LLM) takes over: it reads the selected content and generates a clear, coherent answer in natural language.

By combining semantic search and text generation, RAG offers precise, content-based answers. This approach makes RAG Chatbot a smart, reliable assistant—fully grounded in your own app content.

Setting up RAG Chatbot in your GoodBarber app is simple. You just activate the extension from your back office. You then select the AI model you want to use and customize the chatbot’s appearance to match your app's design—colors, icon, and more.

Next comes the key step: choosing which content to index. RAG Chatbot can draw from your published articles, events, or map points. It works with content from GoodBarber’s built-in CMS, as well as external sources synced with your app, like WordPress, RSS feeds, or custom connectors. Once your selection is made, indexing begins automatically. Your content is analyzed, split into sections, and transformed into vector embeddings, ready to be used by the AI.

From that point on, there’s nothing else to do. Indexing is continuous: every time you publish or update content, RAG Chatbot refreshes its knowledge base to ensure it always provides accurate and up-to-date answers.

In the app, RAG Chatbot appears as a standalone section. You can even add multiple chatbot sections to the same app. This allows you to create specialized assistants for different types of content—for example, one for upcoming events, another for local news, or one dedicated to cultural resources. It’s a simple way to guide users and structure how information is accessed.

RAG Chatbot is especially useful in apps where fast, easy access to information is essential. A tourism app can help users by answering questions like “What’s happening this weekend in Barcelona?” using events pulled directly from the app. A university app can help students find enrollment details or course schedules. A health information app can answer “What are the symptoms of dengue?” by referencing trusted published content. A local government app in Montreal can guide citizens on where to renew a passport. Even a coworking network in Berlin can use RAG Chatbot to centralize internal documentation and give members quick access to guides, rules, or policies.

Whatever the use case, the idea is the same: users ask a question in natural language, and receive a reliable answer based on the actual content of the app. It’s a faster, more intuitive, and more engaging way to access information.

RAG Chatbot works seamlessly with the Memberships extension (In-App Purchase / Subscription). When content is restricted to subscribers, the chatbot respects those rules: it only provides access if the user has the proper rights. This ensures a personalized and secure experience while protecting the value of your premium content.

You can also restrict access to RAG Chatbot sections themselves—limiting them to paying users only. This gives you the flexibility to create exclusive assistants for customer support, premium resources, or expert content. It’s an effective way to increase the perceived value of your subscription offers.