Hermes Agent
The AI Agent That Grows With You
Not a chatbot. Not a copilot. An autonomous agent that lives on your server, remembers everything it learns, builds its own skills, and reaches you on Telegram, Discord, Slack, or WhatsApp — wherever you are.
What Is Hermes Agent?
Hermes Agent is an open-source autonomous AI agent built by Nous Research — the lab behind the Hermes, Nomos, and Psyche model families. Unlike coding copilots tethered to an IDE or chatbot wrappers around a single API, HermesAgent is a self-improving agent that lives on your infrastructure, remembers what it learns across sessions, and gets more capable the longer it runs.
It runs anywhere — a $5 VPS, a GPU cluster, or serverless infrastructure like Daytona or Modal that costs nearly nothing when idle. Talk to it from Telegram while it works on a cloud VM you never SSH into yourself. Hermes Agent is not tied to your laptop. It's your always-on AI that works while you sleep.
Overview
At a Glance
Persistent Memory
Cross-session memory with FTS5 recall and LLM summarization. Hermes Agent remembers your projects, preferences, and context — forever.
Self-Created Skills
The agent autonomously creates procedural skills from experience, improves them during use, and reuses them across sessions. It gets smarter every day.
Lives Where You Do
CLI, Telegram, Discord, Slack, WhatsApp — all from one gateway. Reach your agent from any device, any platform, any time.
Runs Anywhere
Local, Docker, SSH, Daytona, Singularity, Modal. Deploy on a $5 VPS or serverless cloud. Costs nearly nothing when idle.
Core Capabilities
What Hermes Agent Delivers
A closed learning loop, multi-platform presence, and 40+ built-in tools — designed to get more capable the longer it runs.
Closed Learning Loop
Agent-curated memory with periodic nudges, autonomous skill creation, skill self-improvement during use, and Honcho dialectic user modeling. HermesAgent builds a deepening model of who you are.
40+ Built-in Tools
Web search, file operations, terminal commands, image generation, TTS, vision, and more — all out of the box. No configuration required to get started.
Skills System
Procedural memory the agent creates and reuses. Skills are portable, shareable, and community-contributed via the Skills Hub at agentskills.io. Compatible with open standard skills.
MCP Integration
Connect Hermes Agent to any MCP server for extended tool capabilities. Programmatic Tool Calling via execute_code collapses multi-step pipelines into single inference calls.
Scheduled Automations
Built-in cron scheduler with delivery to any platform. Set Hermes Agent to run tasks, send reports, or monitor systems on any schedule — fully autonomous.
Use Cases
Where Hermes Agent Shines
Developer Automation
Spawn isolated subagents for parallel workstreams. Run terminal commands, manage files, search the web, and execute code — all autonomously while you focus on what matters.
Personal AI Assistant
Talk to Hermes Agent on Telegram while it works on a cloud VM. Schedule automations, get reports delivered to any platform, and let it handle recurring tasks on cron.
Research & Analysis
Web search, content extraction, vision, and batch processing built in. Export trajectories for RL training with Atropos. Research-ready from day one.
Team & Enterprise
Connect to Slack or Discord for team-wide AI assistance. MCP integration extends capabilities to any tool. Context files shape every conversation to your project's needs.
Technical Architecture
How Hermes Agent Is Built
Hermes Agent is designed for maximum flexibility and self-sufficiency. It supports 6 terminal backends (local, Docker, SSH, Daytona, Singularity, Modal), enabling deployment from a personal laptop to enterprise serverless infrastructure. The memory system uses FTS5 full-text search with LLM summarization for cross-session recall. The skills system stores procedural memory as portable, shareable skill files compatible with agentskills.io. Built by Nous Research — the lab behind the Hermes model family.
| Terminal Backends | Local, Docker, SSH, Daytona, Singularity, Modal |
| Messaging Platforms | CLI, Telegram, Discord, Slack, WhatsApp |
| Built-in Tools | 40+ (web, file, terminal, vision, TTS, image gen) |
| Memory System | FTS5 cross-session recall + LLM summarization |
| Skills System | Autonomous creation, self-improvement, agentskills.io |
| Model Providers | Nous Portal, OpenRouter, OpenAI, any endpoint |
Why Hermes Agent
Why Choose HermesAgent
Most AI tools are stateless — every conversation starts from zero. Hermes Agent is different. It accumulates knowledge, builds skills, and becomes more useful the longer you use it.
- ✓ Persistent memory across all sessions — Hermes Agent remembers your projects, preferences, and context without you repeating yourself.
- ✓ Self-improving skills — the agent autonomously creates and refines procedural skills from experience, sharing them via the open agentskills.io standard.
- ✓ Truly platform-independent — runs on any infrastructure, talks to you on any messaging platform, works while your laptop is closed.
- ✓ Built by model trainers — created by Nous Research, the lab behind Hermes, Nomos, and Psyche. Works with any OpenAI-compatible endpoint.
Open Source & Access
Free, Open, Self-Hosted
Hermes Agent is fully open source. Install it in 60 seconds on Linux, macOS, or WSL2. Self-host on any infrastructure — from a $5 VPS to enterprise serverless. No vendor lock-in, no usage limits, no subscription required.
Connect to Nous Portal, OpenRouter, OpenAI, or any OpenAI-compatible API endpoint. The agent works with the models you already use. Community skills are available at agentskills.io — portable, shareable, and free.
Release Timeline
Key Milestones
- 2024 — Nous Research releases Hermes 3, establishing the model foundation for Hermes Agent's reasoning and instruction-following capabilities.
- Early 2025 — Hermes Agent launches as an open-source autonomous agent with persistent memory, skills system, and multi-platform messaging gateway.
- Mid 2025 — Skills Hub at agentskills.io goes live; community-contributed skills become portable and shareable across Hermes Agent installations.
- 2025–2026 — Continuous self-improvement loop: agent-curated memory, Honcho user modeling, MCP integration, and Daytona/Modal serverless support added.
Get Started
Install Hermes Agent in 60 Seconds
Install
Run the one-line installer on Linux, macOS, or WSL2. Hermes Agent sets up everything automatically — no manual configuration needed.
Configure
Set your model provider (Nous Portal, OpenRouter, OpenAI, or any endpoint) and connect your messaging platform — Telegram, Discord, Slack, or WhatsApp.
Run
Start your first conversation. Hermes Agent begins learning immediately — building memory, creating skills, and becoming more capable with every session.
Frequently Asked Questions
FAQ
What is Hermes Agent?
Hermes Agent (HermesAgent) is an open-source autonomous AI agent built by Nous Research. It features persistent cross-session memory, a self-improving skills system, multi-platform messaging (Telegram, Discord, Slack, WhatsApp), 40+ built-in tools, and runs on any infrastructure from a $5 VPS to serverless cloud.
How is Hermes Agent different from ChatGPT or Claude?
ChatGPT and Claude are stateless — every conversation starts fresh. Hermes Agent maintains persistent memory across all sessions, autonomously creates and improves skills from experience, runs scheduled automations, and operates independently on your own infrastructure. It gets smarter the longer you use it.
Who built Hermes Agent?
Hermes Agent is built by Nous Research, the AI lab behind the Hermes, Nomos, and Psyche model families. Nous Research is known for producing some of the most capable open-source language models available.
What messaging platforms does Hermes Agent support?
Hermes Agent supports CLI, Telegram, Discord, Slack, and WhatsApp — all from a single unified gateway. You can switch between platforms seamlessly while the agent maintains full context.
Is Hermes Agent free and open source?
Yes. Hermes Agent is fully open source and free to self-host. Install it on any Linux, macOS, or WSL2 system in 60 seconds. No subscription, no usage limits, no vendor lock-in.
What AI models does Hermes Agent work with?
Hermes Agent works with Nous Portal, OpenRouter, OpenAI, and any OpenAI-compatible API endpoint. You can use the Hermes model family or any other model you prefer.
Start Using Hermes Agent Today
Install in 60 seconds on Linux, macOS, or WSL2. Connect to Telegram, Discord, Slack, or WhatsApp. Let Hermes Agent start learning, building skills, and working for you — autonomously.
Get Started