Hermes

High-performance AI inference runtime

Connect Hermes to ONE's decentralized GPU network via the OpenAI-compatible API. Run fast, uncensored models on community hardware using your $ONE stake.

What the integration will enable

—Route Hermes model requests through ONE's inference network
—Use $ONE stake credits for free daily inference
—Access the full ONE model catalog from Hermes
—Dedicated lane for production-grade throughput

Manual setup

works today

ONE speaks the OpenAI-compatible API — you can connect any tool that supports a custom base URL right now.

1Set ONE as your OpenAI-compatible base URL

# In your Hermes config or environment
OPENAI_API_BASE=https://api.onecompute.xyz/v1
OPENAI_API_KEY=<your ONE session key>

2Or pass as runtime flags

hermes \
  --api-base https://api.onecompute.xyz/v1 \
  --api-key <your ONE session key> \
  --model one/claude-sonnet

Official integration in progress

The first-class Hermes plugin is being built. Follow @onecomputexyz for updates.