← All integrations

Standard Compute launcher · Browser sign-in · Personal agent

Use Standard Compute with Hermes Agent

Nous Research’s open-source self-improving agent with persistent memory. Install the Standard Compute launcher, approve the browser sign-in, and Hermes Agent runs on your monthly compute budget.

Without the launcher, run hermes setup, choose Full setup and pick Custom OpenAI-compatible endpoint: base URL https://api.stdcmpt.com/v1, your Standard Compute key and the model standardcompute.

Eligible new accounts can try $0.25 of compute, with no card. See trial terms →

Connect and launch

One browser sign-in covers every agent the launcher supports.

1. Install once

Install Hermes Agent first if needed: Hermes installation guide. Then install or update the SC launcher below.

On Windows, run these commands inside WSL.

curl -fsSL https://standardcompute.com/cli/install.sh | sh

Open a new terminal after installation. Requires Node.js 20 or newer.

2. Launch from your project folder

standardcompute launch hermes

Choose an API key and approve the connection in your browser when prompted. Hermes Agent opens automatically.

Use this command each time. We’ll remember your selected key.

Check the connection

Ask it to reply “OK”, then check Request Logs for the successful request and model used.

Uses a separate, saved SC profile with its own sessions. Your usual setup and connected messaging services are preserved. Run hermes normally to return to them.

Requests use your Standard Compute budget.

Setup details for Hermes Agent

  1. Install Hermes Agent, then install the Standard Compute launcher.
  2. Run standardcompute launch hermes and approve the browser connection.
  3. Hermes Agent runs on a saved Standard Compute profile with its own sessions; your usual Hermes Agent profile is preserved.

Full in-dashboard guide → · Hermes Agent docs ↗

What changes in Hermes Agent

Billing
Hermes Agent requests draw on one monthly compute budget, with plans from $19 a month. When the budget is used, requests stop until renewal or an upgrade, unless you turn on pay as you go.
Models
The model standardcompute picks a frontier or an efficient model for each request. An exact model ID from the Models page pins one model instead.
Good to know
Terminal setup for macOS, Linux, and WSL. External tools may need separate credentials.

Read next