Zola

Zola is an open-source, self-hosted web app that provides a unified web interface that allows users to interact with various large language models through a single application, removing the need to switch between different provider websites. This tool is designed for individuals and teams who want a centralized hub for their AI interactions while maintaining control over their data and API usage.

The software is deployed as a web application and can be run locally via npm or through Docker containers. It supports both cloud-based models via API keys and local models hosted on the user's own hardware, making it suitable for both cloud-reliant workflows and air-gapped or private environments. The deployment process is designed to be straightforward, with specific configurations available for OpenAI cloud integration or local Ollama setups.

Key features

• Support for OpenAI, Mistral, and Ollama models
• Integration with OpenRouter for bring your own API key functionality
• Local model execution and automatic detection via Ollama
• File upload capabilities for model interaction
• Customizable user system prompts and layout options
• Responsive user interface with light and dark theme support
• Integration with Supabase for authentication and storage
• Model integration handled through the Vercel AI SDK

The application is built using a modern stack consisting of Next.js, TypeScript, and Tailwind CSS. It utilizes shadcn/ui for its core components and motion-primitives for animated elements. For users prioritizing privacy or offline access, the integration with Ollama enables the execution of models on local infrastructure without sending data to external servers. The architecture also incorporates prompt-kit to manage AI-specific components and is currently working toward full Model Context Protocol support.

Zola serves as a flexible frontend for users who manage multiple AI providers or prefer to host their own model interfaces to avoid vendor lock-in.