FreX By OpenClaw · Est. 2026
§ Open source · Nous Research

Meet Hermes — the self-improving agent.

A fully open-source agent that learns from every task. It writes its own skills, schedules its own work, and keeps getting better the longer you run it.

By Nous Research · OpenClaw-compatible · WSL2 + Linux + macOS

License
MIT — Free forever
Platforms
macOS · Linux · WSL2
Runtime
Python 3.11+
Setup time
~20 minutes
§ 01 — What it does

Three things, very well.

Core capabilities
CAP. 01
Built-in learning loop
Every completed task feeds back into the agent's skill library. It notices patterns, extracts heuristics, and applies them next time. No training runs — pure in-context learning on your own data.
CAP. 02
Autonomous skill creation
When Hermes finishes a complex multi-step task, it writes a reusable skill for next time. You come back a month later and discover it has built a personal toolkit you didn't know existed.
CAP. 03
Scheduled automation
Tell it "every morning at 9, summarize my email and draft replies." It runs forever. Cron in natural language, backed by persistent state so it actually remembers what it did yesterday.
§ 02 — Hermes vs. OpenClaw

When to pick which.

Honest comparison
AxisHermes AgentOpenClaw
CostFree — MIT licenseFree core + $15 guide / subs
Best forResearchers, tinkerers, Linux-nativesBuilders who want it running today
LearningSelf-improving by defaultMemory layers + manual skill install
MessagingVia plugins — some assembly required6 channels out of the box
Setup~20 min, Linux/WSL2 preferred~30 min, all platforms
PhilosophyOpen weights, open everythingBring your own model, bring your own stack
Windows note

Hermes requires WSL2 (Windows Subsystem for Linux). Native Windows is not supported. Set it up with wsl --install before continuing.

§ 03 — Installation

Five steps. Twenty minutes.

macOS · Linux · WSL2
STEP 01

Clone the repo

Pull the official Nous Research source. The repo includes installation scripts for macOS and Linux.

bash
git clone https://github.com/nousresearch/hermes-agent
cd hermes-agent
STEP 02

Run the installer

Installs Python dependencies, downloads the base models, and sets up the local vector store for skills.

bash
./install.sh
STEP 03

Configure your provider

Hermes works with any OpenAI-compatible API — OpenRouter, Anthropic, a local Ollama server, or a self-hosted vLLM instance.

bash
cp config.example.yaml config.yaml
# then edit config.yaml with your keys
STEP 04

Start the agent

Launches the Hermes loop. It boots, loads its skill library, and starts listening on port 8080.

bash
hermes start
STEP 05

(Optional) Migrate from OpenClaw

If you already have an OpenClaw setup, Hermes can import your SOUL.md, MEMORY.md, and installed skills automatically.

bash
hermes migrate --from ~/.openclaw

Free. Open. Yours.

Hermes is MIT-licensed and built by Nous Research. Clone it, fork it, modify the loop — no permission required.

Get Hermes on GitHub → Or set up OpenClaw →