Tool setup
Hermes Agent Integration Guide
Point Hermes Agent at SeaLink to access our chat models with a single API Key, and switch models by task whenever you need.
Prerequisites
- Hermes Agent installed (run hermes --version to verify)
- Node.js ≥ 20 (run node -v to check)
- A valid SeaLink API Key (starts with sk-sealink-)
- Network access to sealink.io (port 443, HTTPS)
Setup Steps
Install or update Hermes Agent
Install with: curl -fsSL https://get.hermes.run | bash If you already have it installed, run hermes update to get the latest version.
Edit configuration file
Open your config file: nano ~/.hermes/config.yaml Find the model section and update it to: model: default: qwen3.5-plus provider: sealink base_url: https://test.sealink.io/v1 parameters: temperature: 0.6 max_tokens: 16384
Add SeaLink provider
Add a SeaLink entry to the custom_providers section: - name: sealink base_url: https://test.sealink.io/v1 api_key: sk-sealink-your-key api_mode: chat_completions models: - qwen3.5-plus - deepseek-v4-pro - gpt-4o-mini - claude-opus-4-6 The models list specifies which SeaLink models you want available in Hermes — feel free to add or remove entries.
Get an API Key
Head to your SeaLink dashboard → Keys and create a new key (starts with sk-sealink-). Copy the full key string. Don't have an account yet? Visit /login and sign in with your email — an account will be created automatically on your first login.
Verify the setup
Test with: hermes "Hello, please respond in English" If you get a response, your configuration is working. Run hermes -h to see all available options.
Switch models dynamically
Change the default model by editing your config file: model.default: deepseek-v4-pro Or run hermes config for an interactive setup where you can pick your model directly.
SeaLink Advantages
Major model families
Access Qwen, DeepSeek, Claude, GPT, Gemini, and more
Flexible model switching
Choose different models per task — no need to reconfigure your integration
Usage-based billing
Token consumption and cost tracked per model call
Transparent usage
Review request usage and cost breakdown anytime in the SeaLink dashboard
Configurable rate limits
Independent RPM and TPM limits per API Key
Common Issues
Q: "command not found" when running hermes
A: Hermes Agent isn't installed or the install hasn't taken effect yet. Run: curl -fsSL https://get.hermes.run | bash After installation, close and reopen your terminal (or run source ~/.zshrc / source ~/.bashrc) for the PATH to update.
Q: "401 Unauthorized" authentication failure
A: Let's go through the checklist: 1. Check that api_key in ~/.hermes/config.yaml is correct, with no extra spaces 2. Run echo $OPENAI_API_KEY or check the config file to confirm the key is filled in 3. Verify in your SeaLink dashboard that the key hasn't been disabled or expired
Q: Connection timeout or slow responses
A: Here's what we recommend: 1. Confirm your network can reach https://test.sealink.io 2. Switch to a lightweight model (like gpt-4o-mini) to test basic connectivity 3. Check the SeaLink status page for any ongoing service issues
Q: A specific model is unavailable or returns errors
A: If a particular model isn't working: 1. Check your SeaLink dashboard to confirm the model is included in your plan 2. Verify the model ID spelling in your config's models list 3. Switch to a different model to test whether the issue is model-specific or a configuration problem