# Thunderbolt

An open-source, self-hostable AI chat client that can be deployed on premises across desktop, web and mobile, letting a team pick and switch model providers instead of depending on one vendor.

Thunderbolt is an open-source, cross-platform AI client designed to be deployed on-premises. It targets enterprise customers who want to self-host their AI assistant rather than rely on a single hosted vendor, and it runs on web, iOS, Android, macOS, Linux and Windows.

Its backend can be deployed with Docker, and model providers are configured directly, including Ollama or llama.cpp for local inference, or an API key for any OpenAI-compatible provider. The project is still early: it currently depends on an authentication and search backend, though search can be disabled, and is undergoing a security audit ahead of production readiness.

Website: https://thunderbolt.io