regolo

regolo is an inference API from Seeweb, an Italian hosting provider, serving open-weight models from European data centres. The endpoint is OpenAI-compatible, so existing clients work by changing the base URL.

The catalogue covers Mistral Small, Qwen3.5, GPT-OSS, Gemma and Llama models, plus Whisper for transcription. Beyond the shared API, you can deploy a custom model from Hugging Face onto rented GPUs, billed hourly.

Data residency is the pitch: EU-only processing, zero data retention, and infrastructure certifications including ISO 27001 and CISPE. Useful if GDPR or jurisdiction rules out US-hosted providers.