Hermes Agent ☤
Use Deplora's Web3-native AI models in Hermes Agent for autonomous blockchain-powered development.
Hermes Agent is an open-source self-improving AI agent framework by Nous Research. It features persistent cross-session memory, a built-in learning loop, 40+ integrated tools, and multi-platform access (CLI, Telegram, Discord, Slack, WhatsApp).
Prerequisites
- A Deplora API Key with an active subscription
- A computer with terminal access (macOS, Linux, Windows with WSL2, or Android via Termux)
1. Installing Hermes Agent
Hermes Agent offers multiple installation paths depending on your operating system.
macOS & Linux
The standard installation uses a one-liner that handles Python, Node, and all dependencies automatically:
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/install.sh | bash
Verify the installation:
hermes doctor
Windows (Native Beta)
irm https://raw.githubusercontent.com/NousResearch/hermes-agent/main/install.ps1 | iex
For the most battle-tested Windows experience, using the macOS/Linux installer inside WSL2 is recommended.
Android / Termux
pkg update
pkg install -y git python clang rust make pkg-config libffi openssl
git clone --recurse-submodules https://github.com/NousResearch/hermes-agent.git
cd hermes-agent
python -m venv venv
source venv/bin/activate
export ANDROID_API_LEVEL="$(getprop ro.build.version.sdk)"
python -m pip install --upgrade pip setuptools wheel
python -m pip install -e '.[termux]' -c constraints-termux.txt
2. Configuring Deplora as the LLM Provider
Deplora integrates as a Custom Endpoint inside Hermes Agent's provider system.
Open your terminal and start the Hermes model configuration:
hermes setup model
Navigate the Inference Provider menu and select:
(+) Custom endpoint (enter URL manually)
Fill in the provider fields:
| Field | Value |
|---|---|
| API Base URL | https://api.deplora.ai/v1 |
| API Key | Your Deplora API Key |
| Compatibility Mode | 1. Auto-detect [current] (or Chat Completions if auto-detect fails) |
| Model | Leave blank for auto-detect, or specify deplora-one / deplora-core |
Give the connection a display name — e.g., Deplora.
Verifying the Connection
Once configured, launch a chat session or run the Hermes doctor to confirm the endpoint is reachable:
hermes doctor
The wizard automatically maps standard header flows, but remember Deplora natively uses the x-api-key header for authentication.
3. Start Using Hermes Agent with Deplora
Run hermes to start a session powered by Deplora's Web3-native AI models:
hermes
Hermes Agent's persistent memory and self-improving skills system means it gets better the more you use it — your Web3 development patterns, project context, and preferences are remembered across sessions automatically.
Next Steps
Use Deplora directly via HTTP requests.
Try Deplora models in the browser.