Hermes Agent on SovereignEG

Goal: run Hermes Agent (Nous Research's open-source personal agent — terminal chat, tools, memory, and optional Telegram/Discord bots) on SovereignEG models.

Time: 10 minutes.

Hermes works with any OpenAI-compatible endpoint through its Custom endpoint provider. You give it the base URL (https://backend.sovereigneg.com/v1), the key, and a model name.

Step 1 — Install

# Linux / macOS / WSL2
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
source ~/.bashrc      # or ~/.zshrc

# Windows PowerShell
iex (irm https://hermes-agent.nousresearch.com/install.ps1)

Check: hermes --version

Step 2 — Point it at SovereignEG (wizard — easiest)

hermes model

In the wizard choose Custom endpoint (self-hosted / VLLM / etc.) and answer:

API base URL: https://backend.sovereigneg.com/v1
API key:      sk-YOUR_SOVEREIGNEG_KEY
Model name:   SovereignEG/Qwen3.8-27B-FP8

Hermes saves the URL, key, and model in ~/.hermes/config.yaml, which is the single source of truth for model settings.

Step 2 (alternative) — Edit the files by hand

Keep the key out of the yaml: put it in ~/.hermes/.env and point to it with key_env.

~/.hermes/.env:

SOVEREIGNEG_API_KEY=sk-YOUR_SOVEREIGNEG_KEY

~/.hermes/config.yaml:

model:
  default: "SovereignEG/Qwen3.8-27B-FP8"
  provider: "custom"
  base_url: "https://backend.sovereigneg.com/v1"
  key_env: "SOVEREIGNEG_API_KEY"   # or: api_key: "sk-..." inline
  # context_length: 131072      # only if auto-detect fails

Want several named endpoints (for example SovereignEG and a local vLLM)? Use the providers map and pick with custom:<name>:

providers:
  sovereigneg:
    api: "https://backend.sovereigneg.com/v1"                  # `base_url` also works
    key_env: "SOVEREIGNEG_API_KEY"              # variable name in ~/.hermes/.env
    transport: "chat_completions"
 
model:
  default: "SovereignEG/Qwen3.8-27B-FP8"
  provider: "custom:sovereigneg"

Check what Hermes will use:

hermes config get model.default

Step 3 — Chat

hermes            # or: hermes chat

Try: "List the files in this folder and tell me what this project does." Hermes can run shell commands, read and write files, browse, and remember facts between sessions.

Switch models without leaving the chat:

/model custom:sovereigneg:SovereignEG/Qwen3.8-27B-FP8

Or for one session only:

hermes chat --provider custom --model SovereignEG/Qwen3.8-27B-FP8

Step 4 — Optional: Telegram / Discord bot

hermes gateway setup     # wizard: pick Telegram / Discord, paste the bot token
hermes gateway start     # run the bot as a background service

Your bot then answers on SovereignEG models with the same model config.

Which SovereignEG models to use

Hermes needs tool calling for its file and shell tools. Use any model with supports_tools: true in /v1/models.

Troubleshooting

Back to Frameworks & Tools