Agents et multi-agents

Hermes Agent

nousresearch/hermes-agent

A self-improving AI agent built by Nous Research that features a closed learning loop, multi-platform messaging integration, and flexible model provider switching. It is designed to run on low-cost VPS, GPU clusters, or serverless infrastructure.

★ 227,1KÉtoiles
⑂ 44,4KForks
29443Problèmes ouverts
PythonLangue
MITLicence
Q@project.QualityScoreScore éditorial

Captures d'écran du projet

Capture d’écran de Hermes Agent Capture d’écran de Hermes Agent

Vue d’ensemble

Hermes Agent is a Python-based AI agent that creates and improves skills from experience, persists knowledge, searches past conversations, and builds user models across sessions. It offers a terminal interface and a messaging gateway that supports Telegram, Discord, Slack, WhatsApp, Signal, and Email. It supports multiple terminal backends including local, Docker, SSH, Singularity, Modal, Daytona, and Vercel Sandbox, allowing it to run in various environments from a $5 VPS to a GPU cluster.

Fonctionnalités clés

  • Self-improving learning loop with autonomous skill creation
  • Multi-platform messaging gateway (Telegram, Discord, Slack, WhatsApp, Signal, Email)
  • Seven terminal backends including serverless options like Modal and Daytona
  • Model-agnostic with support for Nous Portal, OpenRouter, OpenAI, and custom endpoints
  • Built-in cron scheduler for unattended automations
  • Subagent delegation for parallel workstreams
  • FTS5 session search and Honcho dialectic user modeling
  • Compatibility with the agentskills.io open standard
  • MCP integration for extended capabilities

Prérequis, installation et démarrage rapide

Linux/macOS/WSL2/Termux: `curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash`. Windows (PowerShell): `iex (irm https://hermes-agent.nousresearch.com/install.ps1)`. The installer manages uv, Python 3.11, Node.js, ripgrep, ffmpeg, and Git Bash dependencies.

Utilisation

Start the interactive CLI with `hermes`. Configure the model using `hermes model` or `hermes setup`. Start the messaging gateway with `hermes gateway`. Use slash commands like `/new`, `/model`, `/personality`, and `/skills` during conversations.

Compatibilité des modèles et cas d’usage

Supports any model via Nous Portal, OpenRouter, OpenAI, Anthropic, custom endpoints, and other providers. Models can be switched using `hermes model` or `/model [provider:model]` without code changes.

Licence et notes sur les risques

Licensed under the MIT License.

Editorial verification 2026-08-09: repository URL, owner, description, license and repository statistics were reviewed. License metadata: MIT. README was fetched for the channel draft; re-check repository dependencies, releases and model terms before production use.

Publication et maintenance

Not stated in the repository metadata.

Firecrawl

firecrawl/firecrawl

★ 161,1KTypeScript

LangChain

langchain-ai/langchain

★ 143,6KPython

RAGFlow

infiniflow/ragflow

★ 86,7KGo