Setup tutorial by Sriram

tencent/hy3:free Configuration with Hermes agent.

Create an OpenRouter key, connect Hy3 free to Hermes Agent, and verify one test response.

  • Practical setup
  • Beginner-readable
  • Business workflow focus
Work with Sriram

Build the AI automation your business actually needs.

Send your requirement and get a tailored AI automation blueprint built for your business. No fluff, just what actually works.

Get this built for my business

Use OpenRouter with Tencent Hy3 while the free endpoint is available.

OpenRouter currently lists tencent/hy3:free as a free model with a 262K context window, released July 6, 2026, and marked as going away July 21, 2026. Treat it as a temporary free setup path, not a permanent pricing promise.

Tencent also says Hy3 is available through Tencent Cloud TokenHub and third-party platforms including OpenRouter, with limited-period free access on OpenRouter.

What you are setting up

Hermes AgentThe local agent workspace that runs your prompts, tools, workflows, and dashboard or CLI entrypoint.
OpenRouterThe OpenAI-compatible API gateway used to call hosted models with one API key.
Tencent Hy3 freeThe temporary free model ID: tencent/hy3:free.
API base URLhttps://openrouter.ai/api/v1
Secret storageYour API key should live in .env, Keychain, Credential Manager, or another local secret store. Do not paste it into public docs.

1. Create the OpenRouter API key

  1. Open OpenRouter Keys.
  2. Sign in or create an OpenRouter account.
  3. Click Create API Key.
  4. Give the key a clear name such as Hermes Agent Local.
  5. Copy the key once and store it safely. OpenRouter will not show the full key again later.

Use these values inside Hermes Agent:

Model IDtencent/hy3:free
Base URLhttps://openrouter.ai/api/v1
API styleOpenAI-compatible chat completions.

2. Test OpenRouter before touching Hermes

Run one direct API test first. If this fails, fix the key, model ID, or account access before changing Hermes configuration.

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_OPENROUTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tencent/hy3:free",
    "messages": [
      {
        "role": "user",
        "content": "Hello. Reply with one short sentence."
      }
    ]
  }'

A successful response proves the API key, endpoint, and model route are working outside Hermes.

3. Install Hermes Agent on Windows

Use this path when you have the Hermes Agent zip, folder, or Git repository. Replace HERMES_REPO_URL only if your setup is distributed through Git.

  1. Install Git, Python 3.11 or 3.12, and Node.js LTS.
  2. Open PowerShell as your normal user.
  3. Create a dedicated folder for Hermes.
winget install Git.Git
winget install Python.Python.3.11
winget install OpenJS.NodeJS.LTS

mkdir $HOME\Hermes-Agent
cd $HOME\Hermes-Agent

If you received a zip, extract it into this folder. If you received a Git repo, clone it:

git clone HERMES_REPO_URL .
py -3.11 -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install --upgrade pip
pip install -e .

If PowerShell blocks activation, run this once, then open a new PowerShell window:

Set-ExecutionPolicy -Scope CurrentUser RemoteSigned

If the Hermes package includes a dashboard or web app, install Node dependencies too:

npm install
npm run build

4. Install Hermes Agent on macOS

Use this path on Intel or Apple Silicon Macs. If you already have Python, Git, and Node installed, start from the project folder step.

brew install git python@3.11 node
mkdir -p ~/Hermes-Agent
cd ~/Hermes-Agent

If you received a zip, unzip it into this folder. If you received a Git repo, clone it:

git clone HERMES_REPO_URL .
python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
pip install -e .

If the Hermes package includes a dashboard or web app, install and build the Node side:

npm install
npm run build

5. Add OpenRouter to the Hermes environment

Create or edit .env in the Hermes Agent project root. Keep the real key private and never commit this file.

LLM_PROVIDER=openrouter
OPENROUTER_API_KEY=YOUR_OPENROUTER_KEY
OPENROUTER_BASE_URL=https://openrouter.ai/api/v1
OPENROUTER_MODEL=tencent/hy3:free
HERMES_MODEL=tencent/hy3:free

If Hermes uses YAML instead of environment variables, use the same values in the provider block:

provider: openrouter
base_url: https://openrouter.ai/api/v1
model: tencent/hy3:free
api_key_env: OPENROUTER_API_KEY

Restart Hermes after changing provider settings. Most configuration problems come from using the wrong base URL, a model ID typo, or an API key saved in the wrong shell session.

6. Start Hermes Agent

The exact command depends on your Hermes package. Use the one your package documents, then run the health checks below.

CLI packageTry hermes doctor, hermes doctor --fix, then hermes serve --host 127.0.0.1 --port 9119.
Python entrypointTry python -m hermes or the package-specific launcher from the README.
Node dashboardTry npm start after npm install and npm run build.
# Common local dashboard URL
http://localhost:9119

Use localhost first. Do not expose Hermes publicly until authentication, HTTPS, logging, and secret handling are configured.

7. Run the Hermes smoke test

  1. Confirm Hermes starts without provider errors.
  2. Ask a short prompt from the Hermes chat, CLI, or dashboard.
  3. Confirm the response came from OpenRouter or the selected provider log.
  4. Run one small workflow only after the model test works.
Ask Hermes:
Reply with one sentence and tell me which model/provider you are using.

If Hermes reports missing API key, restart the terminal after editing .env or export the variable in the same session that starts Hermes.

Final checklist

  • Git, Python, and Node are installed.
  • Hermes Agent is extracted or cloned into a dedicated folder.
  • The virtual environment activates correctly.
  • Python dependencies install without errors.
  • Node dependencies build if the dashboard is included.
  • OpenRouter direct curl test returns a response.
  • .env contains OPENROUTER_API_KEY, OPENROUTER_BASE_URL, and OPENROUTER_MODEL.
  • Hermes starts locally and answers a small test prompt.
  • You have a replacement model plan before July 21, 2026, because tencent/hy3:free is temporary.

Troubleshooting

401 or unauthorizedThe OpenRouter key is missing, copied incorrectly, or not loaded by the same process that starts Hermes.
Model not foundCheck the exact model ID: tencent/hy3:free. If the date is after July 21, 2026, choose a different available OpenRouter model.
Connection errorUse the base URL https://openrouter.ai/api/v1, not the model page URL.
PowerShell activation blockedRun Set-ExecutionPolicy -Scope CurrentUser RemoteSigned, then reopen PowerShell.
macOS command not foundOpen a new Terminal window after installing Homebrew, Python, Node, or Hermes CLI tools.

Source links

Sriram Varadarajan Sriram VaradarajanReplies personally

Want Hermes configured for your workflow?

The guide gets Hermes connected to a model. The bigger result comes from wiring the right tools, browser tasks, lead flows, approvals, and handoff steps around it.

Map my automation