AI Agent Portal
llms.txtDeploy anything you build. Your AI coding agent gets a dedicated Ubuntu VPS with full root access, free domain name, and custom domain SSL support. Your agent can install and ship what you need. Works with Cursor, Claude Code, ChatGPT, Codex, and most AI code agents.
API Keys Manage your API keys for programmatic access
Get your API key and start deploying in 60 seconds
Create a free account to generate API keys for programmatic VPS deployment.
Create Free AccountAlready have an account? Log in
Connect Your AI Agent Add this configuration to your AI coding tool
{
"mcpServers": {
"repocloud": {
"url": "https://mcp.repocloud.io/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY_HERE"
}
}
}
}
Add to .cursor/mcp.json in your project root.
claude mcp add repocloud \
--transport streamable-http \
--header "Authorization: Bearer YOUR_API_KEY_HERE" \
https://mcp.repocloud.io/mcp
Run this command in your terminal.
{
"mcpServers": {
"repocloud": {
"type": "streamableHttp",
"url": "https://mcp.repocloud.io/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY_HERE"
}
}
}
}
Add to Claude Desktop settings under Developer → Edit Config.
MCP Server URL: https://mcp.repocloud.io/mcp
Authorization: Bearer YOUR_API_KEY_HERE
Add as a custom MCP integration in ChatGPT settings.
{
"url": "https://mcp.repocloud.io/mcp",
"transport": "streamable-http",
"headers": {
"Authorization": "Bearer YOUR_API_KEY_HERE"
}
}
Generic MCP server configuration for any compatible client.
Capabilities What your AI agent can do with RepoCloud
-
Deploy Ubuntu VPS — Provision a fresh Ubuntu 24.04 server with root SSH access
-
Get SSH Credentials — Retrieve IP address, root password, and SSH keys
-
Manage Power State — Power on, power off, and reboot your servers
-
Resize Plans — Scale up CPU, RAM, and disk to a larger plan
-
Custom Domains — Free subdomain included; add your own domain with automatic SSL
-
Destroy Instances — Tear down servers and release resources when done
REST API Reference Use directly via HTTP — no MCP required
API DocsAll endpoints accept Authorization: Bearer YOUR_API_KEY. Base URL: https://repocloud.io/api/v1
| Method | Endpoint | Description |
|---|---|---|
| POST | /instances |
Deploy a new Ubuntu VPS. Body: {"project_name": "my-app", "flavor": "s-2vcpu-2gb"} |
| GET | /instances |
List all API-created instances with status, IP, and credentials. |
| GET | /instances/{id} |
Get details for a specific instance including SSH credentials. |
| POST | /instances/{id}/actions |
Power on, power off, reboot, or resize. Body: {"action": "power_on"} |
| DELETE | /instances/{id} |
Permanently destroy an instance and release all resources. |
Agent Skill File A portable prompt file your agent can follow without MCP
An Agent Skill is a SKILL.md file that teaches your AI agent a workflow step by step.
Unlike MCP, it requires no server connection — your agent reads it and follows the instructions using
SSH, curl, or any tools it already has. Use it as a standalone alternative or alongside the MCP server.
---
name: RepoCloud VPS Deployment
description: Deploy and manage Ubuntu VPS instances on RepoCloud
version: 1.0.0
trigger_keywords:
- deploy
- host
- VPS
- server
- RepoCloud
---
# RepoCloud VPS Deployment
## What This Does
Provision a fresh Ubuntu 24.04 VPS, SSH in as root,
install your application, and manage the server lifecycle.
## Prerequisites
- A RepoCloud API key (starts with `rcmcp_`)
- Account balance of $4+ to deploy
- Base URL: https://repocloud.io/api/v1
## Step 1: Deploy
```bash
curl -X POST https://repocloud.io/api/v1/instances \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"project_name":"my-app","flavor":"s-2vcpu-2gb"}'
```
Save the `instance_id`, `ipv4`, and `root_password` from
the response.
## Step 2: Wait for Active
Poll until status is `active`:
```bash
curl https://repocloud.io/api/v1/instances/{id} \
-H "Authorization: Bearer YOUR_API_KEY"
```
## Step 3: SSH and Install
```bash
ssh root@{ipv4}
# Use the root_password from Step 1
# Install your application, configure services, etc.
```
## Step 4: Clean Up
```bash
curl -X DELETE \
https://repocloud.io/api/v1/instances/{id} \
-H "Authorization: Bearer YOUR_API_KEY"
```
## Available Tiers
| Flavor | vCPU | RAM | Disk | ~$/month |
|-----------------|------|--------|---------|----------|
| s-1vcpu-2gb | 1 | 2 GB | 30 GB | 6 |
| s-2vcpu-2gb | 2 | 2 GB | 40 GB | 9 |
| s-2vcpu-4gb | 2 | 4 GB | 40 GB | 12 |
| s-4vcpu-8gb | 4 | 8 GB | 80 GB | 24 |
| s-8vcpu-16gb | 8 | 16 GB | 160 GB | 49 |
| s-16vcpu-32gb | 16 | 32 GB | 320 GB | 97 |
## Actions
Power off: `{"action":"power_off"}`
Power on: `{"action":"power_on"}`
Reboot: `{"action":"reboot"}`
Resize: `{"action":"resize","flavor":"s-2vcpu-4gb"}`
POST to `/api/v1/instances/{id}/actions`
Save as SKILL.md in your project. Your AI agent will auto-load it when you ask about deployment.
MCP vs. Skill File Two ways to connect — use either or both
MCP Server
A live connection your AI agent calls in real time. The agent discovers available tools automatically and executes them through the protocol.
- Works with Cursor, Claude Code, Claude Desktop, ChatGPT
- Agent discovers tools automatically
- Requires server connection
Skill File
A prompt file your AI agent reads and follows step by step. No server needed — it uses SSH, curl, and any other tools it already has.
- Works with any AI agent that reads files
- No server or connection required
- Portable across all tools
Documentation Resources for your AI agent and for you
- REST API Reference
- OpenAPI JSON
- MCP Server
- mcp.repocloud.io
- Documentation
- docs.repocloud.io
- AI Context (llms.txt)
- mcp.repocloud.io/llms.txt
- Agent Skill File
- mcp.repocloud.io/SKILL.md