1
Choose your plan
Pick the model, channel, and deployment setup that fits your workflow. Every plan includes a private Mercury environment.
Private Agent Infrastructure
Launch Your Private AI Agent
Your personal AI agent, deployed as a polished control system instead of a self-hosting project. Choose your model, connect Telegram, Discord, or WhatsApp, and go live in minutes. Start today.
Mercury • Your Agent
Online
1) Choose AI Model
🔒 Need local Ollama instead? It’s available for On-Site & Ship-In deployments only.
2) Choose Communication Channel
💬 Additional channels & API integrations available (Slack, Teams, Signal, iMessage, etc.)
3) Choose VPS Size
Monthly Annual 17% OFF
🏢 Need larger servers? Contact us for custom enterprise solutions
⚡ Introductory rates — lock in your price now. Rates are subject to change as we grow.
Your Configuration
Claude Sonnet 4.6
Telegram
Small
Starter Plan
Deploy Now
🔒 Secure checkout 🚫 Cancel anytime ⚡ Live in minutes
⚡ High demand — only 7 deployment slots remaining
How It Works
From checkout to live agent in three clean steps
The heavy lifting stays on our side. You pick the setup you want, we provision and configure Mercury, and your dashboard is ready to use without a self-hosting checklist.
2
We set up your Mercury
We spin up the server, configure Mercury, apply your routing choices, and prep the dashboard. No Docker, no terminal, no provisioning maze.
3
Start using it
Log in, claim skills, rename your server, and start chatting with your agent. Your dashboard becomes the control plane from day one.
The Hard Truth
Why most people never actually
use Mercury
It's not that they don't want an AI agent. It's that self-hosting is a nightmare.
→
🛡️
A private instance, just for you
Your own dedicated agent. No shared infrastructure. Fully isolated, fully yours.
Read more →
⚡
Always online, always ready
No downtime, no restarts. Your agent runs 24/7 so you never have to wait.
Read more →
🖥️
You log in and use it. That's it.
Updates, security, scaling — we handle everything in the background.
Read more →
Everything Your Agent Needs
Shipped, deployed, and running in minutes — not weeks.
AI Agent Chat
Your agent lives in Telegram, WhatsApp, Discord, or Signal. Always on, always responsive.
Multi-Agent Workflows
Spawn specialist agents for research, coding, content — orchestrated automatically.
Custom Skills
Install pre-built skills or write your own. Gmail, Calendar, GitHub, Stripe — one command.
Any Messaging Platform
Connect Telegram, WhatsApp, Discord, Signal, iMessage, or Slack — all from one dashboard.
Trading & Automation Bots
Run Polymarket bots, market scanners, and signal agents autonomously, 24/7.
Privacy-First Deployment
Your agent runs on a dedicated private server. No shared infrastructure. Your data stays yours.
Why Builders Choose Mercury
🔒
Your Data Stays Private
Use Mercury Credits for instant access, or BYOK if you prefer direct control. Your data lives on your server either way.
Read more →
⚡
Deploy in Under 5 Minutes
From zero to a personal AI agent in minutes. No coding, no command line, no DevOps.
Read more →
🛠️
We Handle the Hard Part
We handle setup, config, monitoring, and maintenance for you.
Read more →
Explore All Options
📝 Guides & Insights
The Mercury Blog
Master Agentic Engineering
Explore our guides on deployment, agentic engineering, and building autonomous AI workflows.
Before You Deploy
Frequently Asked Questions
Everything you need to know about Mercury and AI agent hosting
What is Mercury?
Mercury is a fully managed AI agent hosting service. We deploy a private AI agent on your own dedicated server so you get a 24/7 operator without touching a single line of code.
- We handle server setup, software installation, and channel configuration
- Your agent is live and ready to use in about 10–15 minutes
- Connect via Telegram, Discord, or WhatsApp — your choice
- Use Mercury Credits (no API accounts needed) or BYOK for direct provider access
What is an AI agent?
An AI agent doesn't just answer questions — it takes action. Unlike a standard chatbot, your Mercury agent runs autonomously 24/7 on your dedicated server, executing real tasks on your behalf.
- Write, test, and deploy code autonomously
- Research topics and deliver summarized reports
- Send and triage emails on your behalf
- Manage your calendar and schedule meetings
- Execute multi-step automation workflows end-to-end
- Reply to messages in Telegram, Discord, or WhatsApp — 24/7
How much does AI agent hosting cost?
Plans start at $29/month. No contracts. Cancel anytime.
- Cloud Starter — $29/mo: 2 vCPU · 2GB RAM · 40GB SSD · Best for personal use
- Cloud Pro — $49/mo: 3 vCPU · 4GB RAM · 80GB SSD · Power users & small teams
- Cloud Business — $89/mo: 4 vCPU · 8GB RAM · 160GB SSD · Heavy workloads & automation
- On-Site Setup — $3,000 one-time: We set up Mercury on your own hardware
- Ship-In Setup — $1,500 one-time: Ship us your device; we configure & return it
Do I need my own API keys?
No — but you can use them if you want to. Mercury gives you two ways to power your agent, so you can choose what fits your comfort level.
- Mercury Credits (easiest): We handle the AI provider relationship. You buy credits from us, and your agent works immediately — no accounts to create, no API keys to manage.
- BYOK — Bring your own keys, or we set you up: Connect your own API key from Anthropic, OpenAI, Moonshot AI, or any supported provider. Your conversations route directly through your account — we never see them.
- Either way, your agent runs on your own dedicated server — not shared infrastructure
- No usage caps or rate limits imposed by Mercury
- Supports Claude Opus 4.7, ChatGPT-5.4, Kimi K2.5, MiniMax M2.7, and Xiaomi MiMo V2 Pro in cloud, plus Ollama for local on-site and ship-in deployments
How does Mercury work?
Mercury runs as a private AI agent on your server. It connects to your chosen AI models, listens in your preferred channels, and stays available around the clock.
- Runs 24/7 on your dedicated cloud server or local hardware
- Connects to Claude Opus 4.6, ChatGPT-5.4, Kimi K2.5, or any supported model through your API keys
- You message it via Telegram, Discord, or WhatsApp
- Autonomously executes tasks, remembers context, and uses tools
- Spawns sub-agents for complex, multi-step work
Is my data private with Mercury?
Your data stays yours — always. Your agent runs on your own dedicated server. Whether you use Mercury Credits or BYOK, your conversations never pass through shared infrastructure.
- Your API keys stay in your control at all times
- Conversations are processed through your own AI provider accounts
- Every deployment runs on a dedicated VPS — never shared hosting
- We never store your chat logs or training data
- Prefer full self-custody? Choose On-Site or Ship-In — hardware never leaves your hands