Launch pricing: first 3 months discounted for early adopters · Limited spots
Now in beta

Want your own AI agent?
Skip the hassle.

No servers to manage. No code to write. No complicated installations. Just pick your AI model, give it a personality, and start chatting — from your browser or Telegram.

Start for free See pricing

No credit card required · Ready in under 5 minutes · Cancel anytime

H

Built on Hermes Agent MIT

An autonomous agent by Nous Research that grows the longer it runs. 47 built-in tools, persistent memory, auto-generated skills, 14+ messaging platforms, sub-agents, and real sandboxing. We host and manage it so you don't have to.

hermes-agent.nousresearch.com GitHub Docs
Powered by leading AI models via OpenRouter
What Hermes can do

Your agent does the heavy lifting

47 built-in tools, persistent memory, and it gets smarter the more you use it. Zero coding on your end.

14+ platforms

Telegram, Discord, Slack, WhatsApp, Signal, Email, CLI and more. Start on one, pick up on another.

Memory & skills

Learns your projects across sessions. Auto-generates reusable skills from experience. Never forgets.

Sub-agents

Spawn isolated sub-agents for parallel workstreams. Each gets its own terminal, context, and tools.

Any model

Claude, GPT-5, Gemini, Kimi via OpenRouter. Switch anytime. See real costs per conversation.

Real sandboxing

Docker, SSH, Singularity, Modal. Container isolation and namespace hardening built in.

Scheduled automations

Natural language cron for reports, backups, and briefings. Runs unattended through the gateway.

How it works

Up and running in minutes

No terminal. No config files. Just three clicks and you're talking to your agent.

1

Sign up & pick a plan

Create your account. Start free or choose a plan that matches your usage.

2

Configure your agent

Choose models, set personality and tone, enable tools, and connect Telegram.

3

Start working

Chat from the dashboard, talk via Telegram, or use voice. Your agent is always on.

Models

Pick any model. See the real cost.

All models are routed through OpenRouter. Prices are per-conversation estimates based on ~1,500 input + 800 output tokens.

Model Input / 1M tokens Output / 1M tokens Per conversation Context
Claude Opus 4.6 Premium $5.00 $25.00 ~$0.028 1M
Claude Sonnet 4 Reliable $3.00 $15.00 ~$0.017 1M
GPT-5.4 Latest $2.50 $15.00 ~$0.016 1M
GPT-5 Balanced $1.25 $10.00 ~$0.010 400K
Gemini 3 Flash Fast $0.50 $3.00 ~$0.003 1M
Kimi K2.5 Best value $0.38 $1.72 ~$0.002 262K
Light user
~$0.60/mo
10 chats/day · Kimi
Active user
~$3/mo
30 chats/day · Kimi
Power user
~$10/mo
50+ chats/day · Kimi
Same on Claude
~$25/mo
30 chats/day · Opus

1 token ≈ ¾ of a word. Prices via OpenRouter, April 2026. Bring your own API key to bypass plan quotas.

Pricing

Simple, transparent pricing

Pay for hosting + a token quota. Scale when you're ready. Bring your own key to remove limits.

Free Beta
$0/mo
Try it out — no commitment
  • 1 Hermes agent
  • 200K tokens / month (Kimi only)
  • Core tools
  • Web dashboard
  • Telegram / Voice
  • BYOK
Start free
Launch price · 3 months
Starter
$29$19/mo
For personal projects and daily use
  • 1 Hermes agent
  • 1M tokens / month (any model)
  • All tools + Telegram
  • BYOK supported
  • yourname.agentcove.ai
  • Custom domain
Get started
Launch price · 3 months
Pro
$79$49/mo
For creators, founders, and power users
  • 1 Hermes agent
  • 5M tokens / month (any model)
  • All tools + Telegram + Voice
  • BYOK supported
  • Custom domain
  • Priority support
Get started

Overage: $5 per extra 500K tokens on Starter, $5 per extra 1M on Pro. With BYOK, there are no token limits — your usage bills to your own provider account.

FAQ

Common questions

Everything you need to know before getting started.

Not at all. AgentCove is designed for creators, entrepreneurs, and non-technical users. Everything is configured through a visual dashboard — no code, no terminal, no server setup. If you can fill out a form, you can launch an agent.
Think of it as your personal AI assistant that's always on. Unlike ChatGPT, it remembers your projects, learns from past conversations, can browse the web, send messages on Telegram, run scheduled tasks, and get better over time. It lives on a server and works even when you're not watching.
Different models have different strengths and prices. Claude Opus is the most capable but costs more per message. Kimi K2.5 and Gemini Flash are extremely affordable for everyday use. You can switch models anytime from your dashboard — no lock-in.
By default, we provide the API key that powers your agent. If you already have your own key from OpenRouter, Anthropic, or OpenAI, you can plug it in. Your usage then bills directly to your provider with no token limits on your AgentCove plan. The dashboard still tracks your costs either way.
Yes. Hermes supports 14+ platforms including Telegram, Discord, Slack, WhatsApp, Signal, and Email. The setup wizard walks you through connecting your preferred messaging app — usually takes under 2 minutes.
Yes. Your conversations, agent config, and memory files belong to you. We don't train models on your data, don't sell it, and don't share it. When you delete your account, everything is permanently removed. Read our Privacy Policy for details.
Your agent won't stop working. We charge a small overage fee ($5 per extra 500K tokens on Starter, $5 per 1M on Pro). You can also upgrade your plan anytime, or bring your own API key to remove limits entirely.
Yes. No contracts, no lock-in. You can downgrade, pause, or delete your agent at any time from your dashboard settings. The free tier is always available if you just want to keep things light.

Ready to meet your agent?

Sign up in 30 seconds. No credit card. No tech skills required.

Create free account