Nous Research · Open Source

An agent framework for every surface.

Hermes runs the same intelligence across terminal, desktop, messengers, IDEs, and dashboards — with persistent memory, reusable skills, and provider-agnostic model access.

What is
Hermes Agent?

Hermes is an open-source agent framework by Nous Research. It lets the same agent run in your terminal, desktop app, messengers, IDEs, and dashboards. It talks to any LLM provider, remembers who you are across sessions, and learns reusable workflows as skills.

Where a chatbot is a conversation, Hermes is a runtime. It handles tool calls, file system access, web requests, multi-platform messaging, scheduled jobs, and long-running autonomous work — all with a single persistent memory layer and a self-improving skill system.

Open source MIT license Multi-platform Multi-model

Capabilities

One core, many surfaces. Hermes routes the same agent across every platform with consistent tool access.

Desktop & TUI

Native Electron desktop app plus an Ink terminal UI with docked widget apps.

Persistent memory

Remembers preferences, environment facts, and lessons learned across sessions.

Skills system

Reusable procedures saved as skills and loaded into future runs automatically.

Messaging gateway

Telegram, Discord, Slack, WhatsApp, iMessage, Signal, Matrix, Teams, Email, and more.

Extensible

Plugins, MCP servers, webhooks, cron scheduling, skins, and desktop UI plugins.

Provider agnostic

Anthropic, OpenAI, Google, xAI, DeepSeek, local models, OpenRouter, and more.

Tech stack

Built for durability and portability. Every component is designed for long-running agent workloads.

Python

Core runtime, tool layer, skills, integrations, and gateway logic.

Electron

Native desktop surface with streaming chat, plugins, and session history.

SQLite + FTS5

Session store with full-text search across conversation history and memory.

20+LLM providers
12+Messaging platforms
MITLicense
PythonCore runtime