THE PROBLEM
THE SOLUTION
Designing Simple &
Intuitive Chat
UI UX
As a team of expert Bubble developers, we designed this product first and foremost for ourselves.
When we incur something new on Bubble, for example, an unexpected bug, work with a new plugin or new documentation, we have to do our research. But that times time and can get us out of flow.
Our usual default is to ask someone in the team if they have incurred the issue. But someone may be busy and so you never get the answer you need.
With the advances in no code and AI, we wanted to create an intuitive chatbot that supports our Bubble developers and the Bubble community to get what they need.
Our focus for the design was to build on the prominent chat style, and prompt users with some of the most popular Bubble questions.
Using LangChain,
Open AI & Python To
Scrape
Bubble Documentation
& Train The Chatbot
Bubble documentation is a helpful and comprehensive guide to those building with Bubble. It's the backbone of Bubble and we wanted to ensure this was a key part of the chatbot.
To use the documentation, we used our expertise in scraping and indexing allowing the support chatbot to efficiently access a vast repository of knowledge.
We implemented a robust Python-based scraping mechanism, extracting texts and metadata from HTML files hosted on GitHub. The data was then structured into a well-organized index, enabling the chatbot to respond rapidly and accurately to user queries.
By organising the extracted data into a comprehensive index, Bubble developers can now effortlessly find answers to their queries, optimising their app development journey. With this feature, Bubble.io developers gained easy access to a wealth of valuable information, improving their development experience.
By integrating GPT-4 language models and employing advanced search algorithms, our chatbot unpacks complex queries, enabling it to return the most relevant results.
This AI-driven approach gives an intuitive and seamless interaction between developers and the chatbot, ensuring accurate and timely query resolution.
This enables developers to ask any question from bugs to integrations, and get the answer they need within seconds.
Our team expertly deployed the chatbot using Flask and Heroku, harnessing the power of virtual environments and efficient dependency management. This ensured a smooth user experience and the ability to handle numerous concurrent queries. As the Bubble.io developer community continues to grow, our chatbot remains a reliable and indispensable resource, always ready to support developers in their journey.
Given the speed and power of no code, we used Bubble for the front end. Having designed on Figma, we ended our designs were pixel perfect going from Figma to Bubble.
The Bubble documentation is an incredible and comprehensive resource but sometimes lacks detail. Every Bubble query is different and we needed to enhance responses with more than just the information from the documentation.
The Bubble community is passionate and helpful. Sometimes the most helpful information is within the community itself and so we needed to integrate this through the Forum.
To do so, we integrated SERP API so whenever a question was asked to the Bubble Buddy, it would also included the most relevant forum responses.
