Skip to content

geniusrise

Run open LLM, speech-to-text and text-to-speech models on infrastructure you control — a cloud account or your own machines — with one binary and one yaml file.

curl -fsSL https://geniusrise.com/install.sh | sh
geniusrise init
geniusrise plan
geniusrise apply

You get one OpenAI-compatible HTTPS endpoint that:

  • picks the most optimized tested runtime per model (vLLM, llama.cpp, speaches)
  • autoscales with scale-to-zero, inside a hard monthly budget
  • authenticates users with API keys and per-key limits
  • stores nothing about you anywhere except your own infrastructure

Start with the quickstart. The design is documented in architecture and security.