Thox provides an edge AI platform that runs entirely on local hardware, ensuring that sensitive data never leaves the user’s device and meeting HIPAA, GDPR, and ITAR compliance out of the box.
Funding
Funding not disclosed
Founders
Product
Problem
Professionals in healthcare, legal, finance, research and other regulated fields must process sensitive data with AI tools, but cloud‑based services expose that data to external servers and create compliance risks under HIPAA, GDPR, and ITAR.
Solution
Thox delivers an edge AI platform that runs entirely on local hardware, ensuring that all data stays on the user’s workstation or dedicated Nova device. The system combines a hardware‑accelerated inference engine with a hybrid Ollama runtime, providing 60‑100 % faster inference for large language models. Built‑in MagStack clustering lets multiple devices be stacked face‑to‑back to increase RAM and model capacity without any network configuration. The platform includes native vision, OCR, and multilingual analysis for over 32 languages, as well as long context windows of up to 384 K tokens. An OpenAI‑compatible API and developer tools (CLI, Docker, IDE extensions) enable seamless integration into existing workflows while preserving privacy. Compliance‑ready architecture and zero cloud dependencies meet HIPAA, GDPR, and ITAR requirements out of the box.
Target Audience
Primary customers are professionals and enterprises in regulated sectors—such as hospitals, law firms, financial institutions, and research organizations—that require private, high‑performance AI for document analysis, imaging, and code assistance.
Features
- 100 % local processing on Thox Nova edge devices with no cloud telemetry
- MagStack™ clustering that magnetically aligns up to 8 devices to scale RAM and run larger models
- Built‑in vision and OCR pipelines with multilingual support for 32+ languages
- Long context windows (256K–384K tokens) for processing full documents, codebases, or regulatory filings
- OpenAI‑compatible and Ollama APIs, plus CLI and IDE plugins for easy integration
- Hardware‑accelerated inference delivering 60–100 % faster performance on edge CPUs/GPU
- Compliance‑first design meeting HIPAA, GDPR, and ITAR standards by default