Hermes Agent

Hermes Agent is an open-source autonomous agent from Nous Research that runs from a terminal, server, or messaging gateway instead of being limited to an IDE tab. It is designed for long-running personal and developer workflows: it can keep persistent memory, create and reuse skills, schedule tasks, delegate work to subagents, and connect to external tools through MCP. The project’s docs cover CLI usage, configuration, messaging, security, tools, memory, skills, cron scheduling, and development workflows.

Top features

• Terminal-first interface: includes a TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output.
• Messaging gateway: runs across Telegram, Discord, Slack, WhatsApp, Signal, email, Home Assistant, Matrix, Mattermost, SMS, webhooks, and CLI from a shared gateway.
• Persistent memory and session recall: maintains user and project memory across sessions, supports session search, and can summarize prior work for later conversations.
• Skills system: supports reusable procedural skills, user-created skills, and skill improvement during use.
• Scheduled automations: built-in cron scheduling lets the agent run reports, audits, backups, and briefings unattended, then deliver results to a chosen platform.
• Subagents and scripted pipelines: can spawn isolated subagents for parallel workstreams and run Python RPC scripts that call tools without spending the main conversation context on every step.
• Tool and MCP extensibility: ships with a documented toolset system and MCP integration so users can add external servers and domain-specific capabilities.
• Multiple execution backends: supports local, Docker, SSH, Singularity, Modal, Daytona, and Vercel Sandbox terminal backends for local, remote, containerized, or serverless-style environments.

Use cases

• Run a personal agent on a VPS or workstation and interact with it from chat apps or the terminal.
• Maintain long-term context for software projects, recurring operations, and personal workflows.
• Schedule daily reports, weekly audits, reminders, backups, or other unattended jobs.
• Delegate parallel research, coding, or operations tasks to isolated subagents.
• Build reusable skill workflows for tasks the agent should perform repeatedly.
• Connect specialized MCP servers for services, home automation, development tools, or internal APIs.
• Use containerized or remote execution backends when tasks require isolation or access to a specific machine.