Skip to main content

​
SmewAI x Hermes quickstart

​
Use case

This page focuses on one thing: connecting SmewAI’s OpenAI-compatible GPT endpoint to Hermes and verifying a minimal working setup.

​
1. Prerequisites

Before you begin, make sure:
  • you already created an API key in SmewAI Console
  • you know the actual model name Hermes should call
  • you can edit Hermes’ config.yaml
  • you are using the OpenAI-compatible endpoint

​
2. Supported models

This guide currently focuses on GPT model access. Common model names:
Model name
gpt-5.5
gpt-5.4

​
Step 1: Define a custom provider

custom_providers:
  - name: custom-smewai
    base_url: https://smew.ai/v1
    api_key: <your-key>
    api_mode: openai_chat_completions
    models:
      - gpt-5.5
      - gpt-5.4

​
Step 2: Point the default model to that provider

model:
  default: gpt-5.5
  provider: custom-smewai
  base_url: https://smew.ai/v1
  api_key: <your-key>
  api_mode: openai_chat_completions

​
Step 3: Disable Smart Model Routing

smart_model_routing:
  enabled: false

​
4. Minimal working config

model:
  default: gpt-5.5
  provider: custom-smewai
  base_url: https://smew.ai/v1
  api_key: <your-key>
  api_mode: openai_chat_completions

custom_providers:
  - name: custom-smewai
    base_url: https://smew.ai/v1
    api_key: <your-key>
    api_mode: openai_chat_completions
    models:
      - gpt-5.5
      - gpt-5.4

smart_model_routing:
  enabled: false

​
5. Success checklist

  • base_url is https://smew.ai/v1
  • api_mode is openai_chat_completions
  • model.provider points to the provider name defined in custom_providers
  • smart_model_routing.enabled is disabled
  • Hermes can send a message and receive a valid model response

​
6. Common issues

​
Why do I get a 401?

Check:
  • whether the API key is valid
  • whether api_key is set in all required places
  • whether Hermes is really loading the config file you edited

​
Why do I get a 404 or model not found error?

Check:
  • whether the base URL is exactly https://smew.ai/v1
  • whether the model name is actually available to your account
  • whether the default model matches a GPT model Hermes supports

​
7. API verification example

curl -X POST "https://smew.ai/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <your-key>" \
  -d '{"model":"gpt-5.5","messages":[{"role":"user","content":"hi"}]}'
If this request succeeds, Hermes troubleshooting becomes much more direct.