A Privacy-Oriented Multi-Model AI Workspace
Research & Development Project — 2026
Susan AI is an experimental AI workspace developed to explore how multiple Artificial Intelligence providers can be brought together within a single, privacy-oriented and user-controlled environment.
The project focuses on creating a unified interface where users can interact with different AI models while retaining greater control over their API credentials, conversations and workflow.
Research Focus
The project explores several areas of applied Artificial Intelligence and software engineering:
Multi-Model AI Architecture — integrating multiple AI providers through a unified interface.
Bring Your Own Key (BYOK) — allowing users to manage and use their own AI provider credentials.
Privacy-Oriented AI Interaction — exploring approaches that minimise unnecessary server-side persistence of personal data and credentials.
AI-Assisted File Analysis — enabling users to work with documents, images and structured files through AI.
Conversation Management — local storage, import/export and organisation of AI-assisted workflows.
Human–AI Interaction — designing a simpler interface for learning, research, writing, coding and creative exploration.
Cross-Platform AI Applications — extending the same AI workspace from the web environment to desktop applications.
Technical Architecture
Susan AI is developed using a modern web technology stack including:
Next.js · React · TypeScript · AI SDK · Tailwind CSS · Electron
The architecture separates the user interface, AI-provider integration, local data management and desktop packaging layers, allowing the system to evolve as additional models and capabilities are introduced.
Key Features
Multi-Provider Support
A single interface for interacting with different AI model providers.
File Upload & Analysis
Support for AI-assisted interaction with supported documents, images and structured files.
Local Conversation Management
Conversation history can be managed within the user's browser, with import and export capabilities.
Privacy-Focused Credential Management
The system is designed around user-managed API credentials rather than requiring a central account-based API key system.
Responsive Interface
Designed for use across different screen sizes and interaction environments.
Desktop Application
The project can also be packaged as a desktop application using Electron.
Research Perspective
Susan AI is not intended merely as another chatbot interface. It is an exploration of how AI systems can be made more modular, provider-independent and user-controlled.
The project investigates the practical intersection of Artificial Intelligence, Human–Computer Interaction, privacy-oriented software design and multi-model computing.
The broader objective is to understand how emerging AI capabilities can be incorporated into everyday learning, research, productivity and creative workflows without making the user completely dependent on a single AI ecosystem.
Development Status
Status: Active Research & Development
Year: 2026
Developer: Susankar Karmakar
Organization: Sanket Pixel Technologies
Project Repository
Susan AI — GitHub Repository