thunderbolt: a cross-platform AI client for on-prem deployment with support for local and frontier models

What it solves

Thunderbolt is designed to eliminate vendor lock-in and provide users with full control over their AI experience. It allows users to choose their own models and own their data by providing a cross-platform client that can be deployed on-premises.

How it works

It functions as an open-source AI client that connects to various model providers. Users can integrate frontier models via API keys for OpenAI-compatible providers or use local inference engines like Ollama or llama.cpp for free, private local inference.

Who it’s for

It is primarily targeted at enterprise customers who require on-premises deployment for data privacy and security, as well as individuals who want a self-hosted AI interface across desktop and mobile platforms.

Highlights

• Cross-platform availability: Works on web, iOS, Android, Mac, Linux, and Windows.
• On-prem deployment: Supports self-hosting via Docker Compose or Kubernetes.
• Model flexibility: Compatible with frontier, local, and on-prem models.
• Open-source: Licensed under the Mozilla Public License 2.0.