What Is Hermes Agent
Hermes Agent is an open-source (MIT license) agent from Nous Research. Its learning workflow can turn useful task experience into reusable skills. Skill creation and reuse are capabilities to inspect and validate, not a guarantee that each task improves accuracy or reduces model calls.
Capabilities include persistent local state, reusable skills and messaging connectors. For the already source-checked Windows, setup and security overview, see our Hermes Agent profile. Local state does not imply that cloud inference, remote tools or messaging traffic remain on your machine.
Prerequisites
Before you begin, make sure you have the following ready:
- Platform: Linux, supported Apple Silicon macOS, native Windows, or WSL2; check the official prerequisites for your installation method.
- Runtime: the source installer prepares managed dependencies; it is not a standalone Node.js binary.
- Model access: a supported provider account or a configured local endpoint. Tool and model access can have separate charges.
- Optional connectors: configure only the messaging platforms and execution backends you intend to use.
Step 1: Install Hermes Agent
The official command-line installation guide provides this Linux / supported macOS / WSL2 source installer. Review the script and prerequisites before running it; this page has not executed it:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bashVerify the installation by checking the version:
hermes --versionNative Windows installation
The official guide provides this PowerShell route; WSL2 is an alternative, not a requirement:
iex (irm https://hermes-agent.nousresearch.com/install.ps1)Use the installation guide for desktop packages and prerequisites. Docker as a tool-execution backend is distinct from a verified full-agent image; no unverified Docker image command is supplied here.
Step 2: Configure Hermes Agent
Prefer the documented setup flow. The default configuration path is ~/.hermes/config.yaml; avoid copying obsolete provider or security fields.
Choose a provider from your shell
Run hermes model outside an active Hermes chat. This wizard can configure a provider, credentials or OAuth, a custom endpoint and a default model. Follow the choices available for your account.
Nous Portal and MISSINGNO
Nous Research announced MISSINGNO on 2026-10-10 as a coding and agentic-reasoning model, temporarily free on Nous Portal. Token efficiency is a vendor claim; AgentSkillsHub has not independently tested it.
- For the official Portal setup flow, run
hermes setup --portalfrom the shell and complete your own account authorization. - Choose MISSINGNO only if the current Portal list for your account includes it. This page has not logged in or verified account-level availability.
- Once a provider is configured, use
/modelinside a Hermes chat to inspect and switch available configured models. It does not add providers or perform OAuth; leave the session and runhermes modelfor that.
UNKNOWN: MISSINGNO’s original developer provenance, parameter count, base architecture, weight license, independently verified benchmark rank, free-period end date and usage quota. A temporary model offer does not make every Portal tool or the complete agent workflow free.
The Portal guide says its setup also enables the Tool Gateway. Review those external tool routes and account terms. Use the CLI reference for provider-specific and local-endpoint options.
Step 3: Connect a Chat Platform
Hermes supports multiple chat frontends. Here are the two most popular setups plus CLI mode.
Telegram or Discord setup
Use hermes setup and the current messaging-platform instructions to configure the intended connector. Review the bot credentials, allowed users and gateway options. Do not copy a guessed platforms YAML block or expose a port merely because a generic tutorial mentions one.
The CLI reference documents gateway commands. Check hermes gateway --help for the installed version before selecting a service mode.
CLI Mode (for Testing)
The simplest way to test your setup is CLI mode. Just run:
hermes chatThis opens an interactive terminal session where you can send messages and watch the agent respond in real time. CLI mode is ideal for verifying your config before connecting a chat platform.
Step 4: Your First Auto-Generated Skill
Give Hermes a bounded task with clear permission to read the intended files. After a useful result, ask it to capture a reusable procedure when appropriate. A fixed number of tool calls is not a guaranteed skill-creation trigger.
- Start with
hermes chatafter configuring the model. - Use a small, authorized task such as summarizing a local sample README.
- Inspect the result before accepting any generated reusable procedure.
- Review the skill files under
~/.hermes/skills/before reuse.
Follow the existing Hermes skills guide for the skill format and workflow. Do not assume a saved skill guarantees faster, safer or cheaper execution.
Deployment Options
Choose where the agent runs separately from where tools and model inference run. Local, VPS and remote tool backends do not have identical features or cost boundaries. Check the installed version and backend-specific documentation.