Hermes Hermes on RunClaw

An assistant that grows with you. Lightweight, learning-first.

Powered by Hermes from Nous Research. Same channels and model pool as OpenClaw — but lighter on resources, kinder to the budget, with a self-improving learning loop baked in.

See pricing

Compute 2 vCPU · 2 GB RAM · 20 GB SSD

What you get with Hermes

The lighter weight class on RunClaw. Same agent core, smaller footprint, lower price.

Self-improving learning loop

After complex tasks, Hermes writes its own skill documents — the approach, the edge cases, the domain knowledge. Each task makes the next one cheaper.

Subagent parallelism

Delegate by spawning isolated subagents for parallel workstreams — the kind of work that used to need multiple assistants.

Scheduled automations

Built-in cron. Your assistant runs jobs on a schedule and delivers results to any channel.

Lower resource cost

Lighter runtime means a lower price. Same RunClaw managed plumbing, smaller engine — that's the entire $4.2 vs $7.5 story.

Tune it without writing code

Hermes exposes the dials that matter from the RunClaw console — personality, reasoning depth, max turns, session reset, idle timeout.

Same 8+ IM channels as OpenClaw
Same built-in model pool — Kimi K2.6, GLM-4.7 Flash
Plus your own SOUL.md system prompt, loaded on every conversation

Channels Hermes supports

Hermes covers every major IM platform RunClaw exposes — the same set as OpenClaw.

Discord

WeChat

Feishu

WeCom

DingTalk

Hermes from $4.2/mo, ready in 2 minutes

7-day free trial. AI models included. The lighter weight class — pick it for budget-friendly autonomy that learns.