Everything AI

Your cloud, run by your AI. Connect any assistant in a minute.

Hostingsurge speaks MCP. Claude, ChatGPT, Cursor, VS Code, Codex and any MCP client get about 120 tools to create projects, deploy, read logs and manage domains, databases, storage and email — signed in with OAuth, and with a confirmation step before anything destructive.

MCP endpoint: https://hostingsurge.com/mcp — Streamable HTTP, OAuth 2.1 or a scoped API token.

OAuth 2.1 sign-inAbout 120 MCP toolsREST API + OpenAPIllms.txt for agents

Connect in one step.

Clients that support OAuth sign you in with your Hostingsurge account — no token to copy. Everything else uses a scoped API token from Settings → API & MCP keys in the dashboard.

Claude.ai and Claude Desktop

Customize → Connectors → Add custom connector, paste the URL, then sign in to Hostingsurge.

Claude Code

Add the server, then run /mcp and choose Authenticate.

claude mcp add --transport http hostingsurge https://hostingsurge.com/mcp

ChatGPT

Settings → Security and login → Developer mode, create an app with the URL and OAuth, then sign in.

Cursor

Add the server to ~/.cursor/mcp.json and press Connect in Cursor’s MCP settings.

VS Code

Add it to .vscode/mcp.json as an http server; VS Code opens the sign-in for you.

OpenAI Codex CLI

Add the URL with a bearer token taken from an environment variable.

codex mcp add hostingsurge --url https://hostingsurge.com/mcp --bearer-token-env-var HOSTINGSURGE_TOKEN

Windsurf

Add the server URL and an Authorization header to Windsurf’s MCP config.

Any MCP client

Point it at the URL over Streamable HTTP with an Authorization: Bearer token.

Copy and paste

# Claude Code — then run /mcp → Authenticate
claude mcp add --transport http hostingsurge https://hostingsurge.com/mcp

# Cursor — ~/.cursor/mcp.json (signs in with OAuth)
{"mcpServers":{"hostingsurge":{"url":"https://hostingsurge.com/mcp"}}}

# Codex CLI — token from Settings → API & MCP keys
codex mcp add hostingsurge --url https://hostingsurge.com/mcp --bearer-token-env-var HOSTINGSURGE_TOKEN

What your AI can do.

The same actions as the dashboard, the CLI and the REST API — grouped the way the MCP server exposes them.

Projects and deploys

Create projects from Git, a template or an upload; deploy, promote, roll back and cancel.

Logs and debugging

Read build and runtime logs, follow a failing deploy and check uptime incidents.

Domains and SSL

Add domains, check DNS records, set redirects and manage certificates.

Databases

Create PostgreSQL, MySQL, MariaDB, Redis or MongoDB; back up, restore and copy.

Storage

Create S3-compatible buckets and manage their keys.

Email

Set up sending domains and mailboxes on your domain (early access).

WordPress

Update plugins and themes, harden, back up and restore WordPress sites.

Migrations and templates

Import from Vercel, Netlify or cPanel, or start from one of the one-click templates.

Build with AI.

When you do not have code yet, the AI Builder writes it — into a project you can read, publish and keep.

AI Builder

Describe a site or app, refine it in conversation, check each change in a preview and publish to your domain.

From your assistant too

Start, refine and publish Builder projects from Claude, ChatGPT or Cursor over MCP.

Credits that scale

5 free credits a day; Pro from $25 for 100 credits a month, up to 10,000 — cheaper per credit as you grow.

Run your AI apps here.

The building blocks for agents, chat apps and automations, next to the rest of your stack.

Vectors in PostgreSQL

pgvector is available on managed PostgreSQL for embeddings and retrieval.

n8n in one click

Self-hosted n8n for AI workflows and automations, from the template gallery.

Any Docker image

Run AI gateways, chat interfaces or vector databases from a public image, on your private network.

Agent runtimes

Private, always-on AI agents with their own isolated environment — in early access.

No GPUs yet — like Railway and Render. Call hosted model APIs (Anthropic, OpenAI, OpenRouter) from your app; the keys live in your project’s environment variables.

Docs made for agents.

Machine-readable references, so an agent can learn the platform without a human in the loop.

  • /llms.txtA plain-text guide to the platform, the MCP tools and the API, written for language models.
  • OpenAPIThe REST API v1 as an OpenAPI document.
  • MCP guideEvery client, every tool group, and the security model.
  • DevelopersAPI tokens, the hs CLI and Git workflows.

Safe by default.

Your AI acts with the access you grant — and asks before it breaks anything.

OAuth with scopes

Clients sign in through a consent screen and get only the scopes you approve. Revoke them any time.

Confirm before destroying

Destructive tools — deleting a project, restoring over a database — need an explicit confirmation argument.

Tokens per team

API tokens belong to one team, carry only the scopes you pick, can be limited to selected projects and are shown once.

Questions

Which AI assistants work with Hostingsurge?

Any client that speaks MCP over Streamable HTTP: Claude.ai, Claude Desktop, Claude Code, ChatGPT (developer mode), Cursor, VS Code, OpenAI Codex CLI, Windsurf and others.

Does it cost extra?

No. MCP, the API and the CLI are included with every plan. The AI Builder is priced in credits on its own price list.

Is there an assistant inside the dashboard?

Not yet. Today you bring your own assistant through MCP; the dashboard’s Settings → AI page gives you the snippets to connect it.

Can I run language models on Hostingsurge?

Not on GPUs — there are none yet. Small CPU models and any AI app that calls a hosted model API run like every other service.

Plug in your AI.

Create a free account, connect your assistant and ask it to deploy your first project.