Hermes is an open-source AI agent designed for long-horizon autonomous tasks. It maintains a persistent memory database across sessions, schedules tasks with natural language deadlines, reflects on its own performance, and integrates with MCP tools.

How does Hermes differ from OpenClaw?

OpenClaw is a conversational agent — you talk to it in sessions. Hermes is autonomous — you give it a goal and a deadline and it executes across hours or days without prompting. Hermes has a proper memory database rather than markdown files. Most users run both.

Which models work best with Hermes?

Claude Sonnet 4.6 for most tasks (200K context, good reasoning). Claude Opus 4.6 for complex multi-week projects. Claude Haiku 4.5 for simple automation where cost matters. Hermes can auto-escalate to heavier models when task complexity warrants it.

Hermes Agent Hub — Long-Running AI Agent Guides 2026

🪁

Hermes

Long-horizon · Persistent memory · Autonomous tasks · MCP tools

MIT licensed v0.9.3 stable SQLite & PostgreSQL memory MCP tool support Self-improving via reflection

Hermes is an open-source AI agent built for tasks that outlast a single conversation. Where OpenClaw handles one session at a time, Hermes maintains a persistent memory database, schedules autonomous workflows with natural language deadlines, and reflects on past performance to improve future tasks. Give it a goal by Friday — it plans, executes, checks in when needed, and delivers.

Upgrade to v0.9.3 before relying on persistent memory

Versions before v0.9.3 had a silent memory expiry bug — all stored memory appeared intact but was lost after 72 hours. Version v0.9.3 (March 2026) fixes this permanently. Run npm install -g hermes-agent@latest to upgrade.

Guides

⚡

Quick Start — 20 Minutes

Install Hermes, run the setup wizard, start the daemon, pick the right model for long-context tasks, and submit your first scheduled task.

Live

🧠

Persistent Memory Architecture

How Hermes's three-tier memory works: episodic (raw sessions), semantic (compressed facts), procedural (learned patterns). SQLite vs PostgreSQL, compression, retrieval tuning.

Live

🗓

Long-Running Tasks & Scheduling

Task submission, natural language scheduling, TASKS.md format, check-ins via Telegram, autonomous multi-step execution, safety controls, and self-reflection after completion.

Live

🔌

MCP Tool Integration

Connect GitHub, web search, filesystem, Puppeteer, PostgreSQL, and more via Model Context Protocol. Covers v0.9.x adapter and v1.0 native MCP, tool authorisation, and writing custom servers.

Live

⚖️

Hermes vs OpenClaw

Full comparison: memory model, execution style, tool ecosystem, cost per outcome, and the recommended hybrid setup — Hermes for long tasks, OpenClaw for conversations.

Live

At a Glance

License	MIT (fully free)
Install	`npm install -g hermes-agent`
Current version	v0.9.3 (March 2026) — upgrade required for stable memory
Requires	Node.js 22.16+ or Node 24; 500 MB+ disk for memory store
Memory store	SQLite (default, personal) or PostgreSQL (team/multi-machine)
Memory types	Episodic · Semantic · Procedural — three-tier architecture
Scheduling	Natural language: "by friday", "every monday 9am", ISO 8601
Tool support	MCP (adapter in v0.9.x; native in v1.0 — Q2 2026)
Notification channel	Telegram or email for check-ins and completion alerts
Recommended model	Claude Sonnet 4.6 (default) with auto-escalation to Opus
Typical monthly cost	$5–30 depending on task frequency and complexity

Related on This Site

OpenClaw hub — conversational agent with rich skill ecosystem; pairs well with Hermes for day-to-day interaction
SOUL.md & Agent Personas — the workspace file system Hermes extends with TASKS.md and REFLECTIONS.md
Cost Optimisation Guide — model tiering and context strategies that apply to Hermes as well
Weekly News Digest — Hermes release notes and MCP ecosystem updates