hermes-cld 0.1.2

The self-improving AI agent — creates skills from experience, improves them during use, and runs anywhere

Hermes Cloud ☤

Built for projects, not just conversations. Use any model — local, Cloudflare Workers, or your own key.

No login, no account, free to start. Your chats and data stay on your machine.

---

What makes Hermes Cloud different

Project-first architecture — conversations are organized around projects, not a flat chat log. Each project has its own context, tools, and scratchpad.

Serverless reasoning on Cloudflare Workers — run inference on Workers AI without provisioning a GPU. Bring your own Cloudflare token or use the built-in free tier.

Privacy-first by default — your chat history, sessions, and data stay on your machine. No accounts, no telemetry, no cloud sync. When you use a remote model (Workers AI, OpenAI, Anthropic), prompts are sent to that provider for inference — that's inherent to remote inference. Choose a local model (Ollama) to keep everything fully offline.

Free to start, no lock-in — bring your own API key for OpenAI, Anthropic, or any OpenAI-compatible endpoint. Or run a local model through Ollama. Switch any time with hermes model.

Built on Hermes Agent — this project is a fork of Hermes Agent by Nous Research. It inherits the core agent loop, tool system, skills, memory, and gateway, and adds project-first organization, Workers AI integration, and a focus on local-first, no-account usage.

---

Quick Install

Linux, macOS, WSL2, Termux

pip install hermes-cld
hermes

Windows (PowerShell)

pip install hermes-cld
hermes

Requires Python 3.11–3.13. See prerequisites below if you're on a bare system.

After installation:

hermes model      # Pick a model (local, Workers AI, or bring your own key)
hermes            # Start chatting!

---

Getting Started

hermes                  # Interactive CLI — start a conversation
hermes model            # Choose provider and model
hermes project new      # Create a new project
hermes project list     # List all projects
hermes tools            # Configure which tools are enabled
hermes doctor           # Check that everything is set up correctly

---

Choose Your Model

Hermes Cloud works with multiple backends — pick whichever fits your workflow:

Backend	Setup	Cost
Cloudflare Workers AI	hermes model → select Workers AI	Free tier included; bring your own Cloudflare token for higher limits
Ollama (local)	Install Ollama, then hermes model → select Ollama	Free (runs on your hardware)
OpenAI / compatible	Set OPENAI_API_KEY, then hermes model → select OpenAI	Your own API key, your own billing
Anthropic	Set ANTHROPIC_API_KEY, then hermes model	Your own API key

You're never locked in — switch with a single command.

---

Project-First Workflow

hermes project new my-research
hermes project switch my-research
hermes project archive my-research

Each project has its own conversation history, tools, and working directory. Switch between them without cross-contamination.

---

Documentation

Topic	What's Covered
Getting Started	Install, setup, first conversation
CLI Commands	Full command reference
Providers & Models	Workers AI, Ollama, OpenAI, Anthropic
Project System	Project management
Configuration	Config file reference

---

Contributing

PRs and issues welcome at github.com/rohandalvi369/Hermes-Cloud.

Development setup:

git clone https://github.com/rohandalvi369/Hermes-Cloud.git
cd Hermes-Cloud
pip install -e ".[dev]"
scripts/run_tests.sh

---

License & Credits

MIT — see LICENSE.

Hermes Cloud is a fork of Hermes Agent by Nous Research, used under the MIT License. See NOTICE for attribution.

The upstream Hermes Agent project is the foundation of this work — its team built the agent loop, tool system, gateway, and learning infrastructure that Hermes Cloud builds on.