Memorybase
This SaaS solution is a custom project I developed with a partner to address a common issue when creating automations with LLMs. It enables you to quickly provide detailed context for AI chatbots, making them accessible across teams and platforms via API key integration.
Role:
Product/Brand Designer
Date:
June 2024
Duration:
3-4 Weeks
Tools:
Figma, linear.app, relume.io, bubble.io
Project:
Memorybase.io
Problem
Achieving high-quality AI output requires a detailed context setup for each chatbot. With multiple topics and tasks, repeatedly explaining context can be time-consuming—especially when you need a centralized knowledgebase for teams or automations.
Solution
A centralised context database that allows you to edit, share and access context across different chatbots and teams.
Process
Given the project’s scope and tight deadline, I rapidly developed wireframes, iterating based on feedback from potential users. This approach enabled a high-fidelity prototype in a short time.
Branding & Design
With only a few weeks to execute the project, I focused on essential design elements to ensure an intuitive MVP. I utilized the Relume Library Figma Kit as a foundational structure, maintaining a minimalistic approach.
Main features
Create a knowledgebase
Easily set up individual knowledgebases with AI support. Define specific topics to enhance your database.
API keys & API snippets
Use your knowledgebase with API keys. Predefined snippets demonstrate how to implement the API without needing to search through documentation.
Chat with knowledge
Access AI models with direct chat functionality, allowing team-wide contextual use across projects.
Further reflections
Memorybase was developed in just 3-4 weeks to solve a specific problem that exists in my immediate environment: providing detailed context for AI chatbots across teams and platforms. Therefore, the project is very much focussed on the user and their actual needs. I used a lean design approach with rapid prototyping and solicited feedback to design the user experience effectively. Unlike larger projects where a little more time is available, I had to prioritise and design the most important features first.
The minimalist design and use of a Figma library helped to quickly create a functional MVP and the API integration allows for easy scalability across different systems.
Possible next steps could be adding an interface for even easier API management to give users more control. However, these steps will only be taken after further testing to ensure the benefits of customisation.