# NoctHost (remote · mcp.nocthost.com)

Create, list and destroy hourly-billed servers on Vultr, DigitalOcean, Hetzner and Linode.

- Trust score: 84/100 (high trust)
- Change this week: +4
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `mcp.nocthost.com`: 84/100 (this document), [markdown](https://verifymcp.io/servers/com-nocthost-mcp/mcp.md), [page](https://verifymcp.io/servers/com-nocthost-mcp/mcp)

## Channel facts

- Endpoint: `https://mcp.nocthost.com/mcp`
- Transports: `streamable-http`
- Auth: `required`
- Version: `2.0.0`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Endpoint Security**: 92/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - Authorisation is enforced on tool calls, advertised via RFC 9728 protected-resource metadata. Discovery is public, which costs nothing: no tool can be invoked without a token.
  - HTTPS is enforced; there's no plaintext access path.
  - HSTS check failed: the Strict-Transport-Security header is absent.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
  - The authorisation server supports Client ID Metadata Documents, the current MCP client-registration mechanism.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 81/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 4246 tokens (~117/item across 36 items; 35 tools + 1 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 33/100
  - Stability observed for 10 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 95/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 85% of tool parameters carry a description.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - All 4 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation.
  - An AI judge read all 37 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the NoctHost MCP server?

NoctHost is a hosted endpoint at https://mcp.nocthost.com/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http com-nocthost-mcp 'https://mcp.nocthost.com/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "com-nocthost-mcp": {
      "url": "https://mcp.nocthost.com/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "com-nocthost-mcp": {
      "type": "http",
      "url": "https://mcp.nocthost.com/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.com-nocthost-mcp]
url = "https://mcp.nocthost.com/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "com-nocthost-mcp": {
      "type": "remote",
      "url": "https://mcp.nocthost.com/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add com-nocthost-mcp --url 'https://mcp.nocthost.com/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  com-nocthost-mcp:
    url: "https://mcp.nocthost.com/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "com-nocthost-mcp": {
      "Transport": "http",
      "Url": "https://mcp.nocthost.com/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add com-nocthost-mcp -t streamable-http -u 'https://mcp.nocthost.com/mcp'
```

### Other

```json
{
  "mcpServers": {
    "com-nocthost-mcp": {
      "type": "http",
      "url": "https://mcp.nocthost.com/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-21 (score 84, +1)

No change was recorded against any check on this day. Stability & Change Management went from 30 to 33. That category is still filling its 30-day observation window: 9 days of observed history at the previous scan, 10 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-19 (score 83, +1)

- [security] The server rewrote its instructions, which are the text every model session reads
- [security] Tool “list_options” rewrote its description, which is the text the model reads

### 2026-11-17 (score 82, +1)

- [security] The server rewrote its instructions, which are the text every model session reads
- [security] Tool “list_options” rewrote its description, which is the text the model reads

### 2026-11-15 (score 81, +1)

No change was recorded against any check on this day. Stability & Change Management went from 10 to 13. That category is still filling its 30-day observation window: 3 days of observed history at the previous scan, 4 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-13 (score 80, 0)

- [security] Tool “create_backup” rewrote its description, which is the text the model reads
- [security] Tool “create_load_balancer” rewrote its description, which is the text the model reads
- [security] Tool “create_server” rewrote its description, which is the text the model reads
- [security] Tool “get_account” rewrote its description, which is the text the model reads
- [security] Tool “list_apps” rewrote its description, which is the text the model reads
- [security] Tool “list_options” rewrote its description, which is the text the model reads
- [security] Tool “list_payment_methods” rewrote its description, which is the text the model reads
- [security] Tool “reinstall_server” rewrote its description, which is the text the model reads
- [security] Tool “reset_root_password” rewrote its description, which is the text the model reads
- [security] Tool “server_power” rewrote its description, which is the text the model reads
- [security] Tool “support_send” rewrote its description, which is the text the model reads
- [security] Tool “top_up” rewrote its description, which is the text the model reads
- [cosmetic] “list_options” added an optional parameter “gpu_model”
- [cosmetic] “list_options” added an optional parameter “min_vram_gb”
- [cosmetic] “change_ip” reworded the description of “confirm”
- [cosmetic] “change_ip” reworded the description of “id”
- [cosmetic] “console_link” reworded the description of “id”
- [cosmetic] “create_backup” reworded the description of “server_id”
- [cosmetic] “create_server” reworded the description of “app”
- [cosmetic] “create_server” reworded the description of “plan”
- [cosmetic] “delete_backup” reworded the description of “confirm”
- [cosmetic] “destroy_server” reworded the description of “confirm”
- [cosmetic] “get_app_password” reworded the description of “id”
- [cosmetic] “get_bandwidth” reworded the description of “id”
- [cosmetic] “get_server” reworded the description of “id”
- [cosmetic] “get_windows_password” reworded the description of “id”
- [cosmetic] “quote” reworded the description of “plan”
- [cosmetic] “reinstall_server” reworded the description of “confirm”
- [cosmetic] “reinstall_server” reworded the description of “id”
- [cosmetic] “rename_server” reworded the description of “id”
- [cosmetic] “reset_root_password” reworded the description of “id”
- [cosmetic] “restore_backup” reworded the description of “backup_id”
- [cosmetic] “restore_backup” reworded the description of “confirm”
- [cosmetic] “set_rdns” reworded the description of “id”
- [cosmetic] “top_up” reworded the description of “method”
- [cosmetic] “top_up” reworded the description of “sanctions_confirmed”

### 2026-11-12 (score 80, +5)

- [security improvement] Authorization: fail → pass
- [security] First check of Authorization: pass
- [security] The server rewrote its instructions, which are the text every model session reads
- [security] New tool “change_ip”, which the server declares destructive
- [security] New tool “delete_backup”, which the server declares destructive
- [security] New tool “delete_load_balancer”, which the server declares destructive
- [security] New tool “reinstall_server”, which the server declares destructive
- [security] Tool “create_server” rewrote its description, which is the text the model reads
- [functional improvement] Schema quality: 152 → 118
- [functional improvement] Tool coverage: 80% → 85%
- [functional improvement] Stability: unverified → 0.03
- [functional] New prompt “get-help-from-support”
- [functional] New prompt “pick-a-plan”
- [functional] Server version: 1.0.0 → 2.0.0
- [functional] New tool “billing_history”
- [functional] New tool “console_link”
- [functional] New tool “create_backup”
- [functional] New tool “create_load_balancer”
- [functional] New tool “get_app_password”
- [functional] New tool “get_bandwidth”
- [functional] New tool “get_windows_password”
- [functional] New tool “list_apps”
- [functional] New tool “list_backups”
- [functional] New tool “list_deposits”
- [functional] New tool “list_invoices”
- [functional] New tool “list_load_balancers”
- [functional] New tool “list_payment_methods”
- [functional] New tool “recommend_plan”
- [functional] New tool “rename_server”
- [functional] New tool “request_server_limit”
- [functional] New tool “reset_root_password”
- [functional] New tool “restore_backup”
- [functional] New tool “set_load_balancer_backends”
- [functional] New tool “set_rdns”
- [functional] New tool “support_read”
- [functional] New tool “support_send”
- [functional] New tool “top_up”
- [cosmetic] “create_server” added an optional parameter “app”

### 2026-11-11 (score 75)

First indexed and scored.

## MCP tools (35)

### `get_account` (~36 tokens)

Account balance and servers

Balance in USD, current hourly spend, how long the balance lasts at that rate, and how many servers the account runs and may run.

### `list_options` (~375 tokens)

Server sizes, prices and cities

What can be deployed right now: sizes (shared vCPU, dedicated CPU, NVIDIA GPU, bare metal) with vCPU, RAM, disk, included traffic, hourly and monthly price in USD, and the clouds (Hetzner, Vultr, DigitalOcean, Linode and OVHcloud, plus bare metal on Cherry Servers) and city ids each one deploys in. Sorted by hourly price. Filter by city, country or continent, cloud, kind, GPU model and minimum VRAM, minimum vCPU/RAM/disk and maximum hourly price. GPU sizes carry gpu_model, gpu_count and vram_gb fields. Paginated; also returns the OS images.

Input parameters:

- `gpu` (boolean): true = only GPU servers, false = no GPU servers.
- `gpu_model` (string): NVIDIA GPU model, e.g. "L40S", "A16", "RTX 4000 Ada", "H100". Implies GPU servers only.
- `kind` (string): vps = shared vCPU, dedicated = dedicated CPU cores, gpu = NVIDIA GPU, metal = bare metal.
- `limit` (integer): Sizes per page, default 15.
- `max_hourly_usd` (number): Only sizes at or below this price per hour in some city.
- `min_disk_gb` (number)
- `min_ram_gb` (number)
- `min_vcpu` (integer)
- `min_vram_gb` (number): Minimum total GPU memory in GB. Implies GPU servers only.
- `offset` (integer): Skip this many sizes (from next_offset).
- `provider` (string): Only this cloud.
- `region` (string): City id or name ("fra", "Frankfurt"), country ("Germany", "US") or continent ("Europe").

### `quote` (~190 tokens)

Price quote

Price of a specific order: plan in a city (and optionally cloud, OS, count, hours). Returns the hourly price per server and in total, the cost for the given hours, the monthly price, and whether the current balance covers the first hour. Without a cloud it picks the cheapest one that deploys there. Does not create anything.

Input parameters:

- `count` (integer): Number of servers, default 1.
- `hours` (number): How long they run, default 1 hour.
- `os` (string): OS id or name, default "ubuntu-24.04". Windows changes the price.
- `plan` (string, required): Plan id, e.g. "vc2-1c-1gb", or a plan name like "Micro".
- `provider` (string)
- `region` (string, required): City id or name, e.g. "fra" or "Frankfurt".

### `list_apps` (~57 tokens)

One-click apps

One-click apps this account can deploy as the app of a new server: id, name, what it is, the minimum RAM and disk, and the clouds it deploys on. Each installs on Ubuntu 24.04 on first boot.

### `list_servers` (~44 tokens)

List servers

Servers on the account (everything not destroyed): id, label, status, cloud, city, plan, IPv4, hourly price and a ready SSH command for running Linux servers.

### `get_server` (~64 tokens)

Server details

One server by id: status (PROVISIONING, ACTIVE, STOPPED, ERROR and so on), IPv4, SSH command, plan, city, cloud and hourly price. Use it to follow a server that was still provisioning.

Input parameters:

- `id` (string, required): Server id.

### `create_server` (~380 tokens)

Create server

Creates 1 to 10 servers (default limit 3 per call) and waits up to about two minutes for ACTIVE, then returns ids, IPv4 and the SSH command. Spends money: billed per hour from the balance until the server is destroyed. Refused with the price if the total hourly price is above max_hourly_usd (default $1); more than 3 servers also need an explicit max_hourly_usd. Without ssh_public_key a random root password is generated and returned once, and it stays in the chat history. app installs a one-click app (Docker, WireGuard, WordPress and others) on Ubuntu 24.04 instead of a plain OS.

Input parameters:

- `app` (string): One-click app id, e.g. "docker" or "wireguard". Replaces os; the app installs on first boot.
- `count` (integer): How many identical servers, default 1.
- `label` (string): Server name: Latin letters, digits, dot, dash, underscore. With count > 1 it gets -1, -2, ...
- `max_hourly_usd` (number): Upper bound for the total hourly price of this call, default 1.
- `os` (string): OS id or name, default "ubuntu-24.04". Ignored for GPU plans.
- `plan` (string, required): Plan id, e.g. "vc2-1c-1gb", or a plan name like "Micro".
- `provider` (string): Cloud; default is the cheapest that deploys this plan in this city.
- `region` (string, required): City id or name, e.g. "fra" or "Frankfurt".
- `ssh_public_key` (string): One public key line (ssh-ed25519 ..., ssh-rsa ...), installed for root.

### `server_power` (~59 tokens)

Reboot, stop or start a server

reboot, stop (power off) or start a server by id. A stopped server keeps its IP and disk and is still billed per hour while it exists; destroying it ends billing.

Input parameters:

- `action` (string, required)
- `id` (string, required)

### `destroy_server` (~90 tokens)

Destroy server

Permanently deletes a server and its disk; billing stops in the hour it is destroyed. Needs confirm: true and the server's exact label. Without both it only describes what would be destroyed.

Input parameters:

- `confirm` (boolean): true destroys the server; without it nothing is destroyed.
- `id` (string, required): Server id.
- `label` (string): The server's exact label, as a second check.

### `recommend_plan` (~295 tokens)

Recommend a server size

Picks the cheapest sizes that fit a workload, across clouds and cities, from what can be deployed now. Describe the workload (e.g. "Minecraft for 12 players", "Llama 3.1 8B with Ollama", "WireGuard VPN for 5 people") and add numbers when known. Returns fitting options with hourly and monthly price, the sizing basis, and for cheaper sizes that do not fit, one line why. Sizing comes from NoctHost's published guides. Does not create anything.

Input parameters:

- `city` (string): City, country or continent, e.g. "Frankfurt", "Germany", "Europe".
- `limit` (integer): How many fitting options, default 5.
- `max_hourly_usd` (number): Budget per server per hour, USD.
- `model_size_b` (number): LLM size in billions of parameters, e.g. 8 for an 8B model (4-bit quantized).
- `monthly_traffic_gb` (number): Expected outbound traffic per month, GB.
- `need_gpu` (boolean): Only NVIDIA GPU servers.
- `need_windows` (boolean): Windows Server instead of Linux (its own price, at least 4 GB RAM).
- `users` (integer): Players, VPN users, streams or services, depending on the workload.
- `workload` (string, required): What the server is for, in plain words.

### `rename_server` (~55 tokens)

Rename server

Changes a server's label (also renamed at the cloud). Latin letters, digits, dot, dash and underscore, up to 64 characters.

Input parameters:

- `id` (string, required): Server id.
- `label` (string, required): New label.

### `set_rdns` (~85 tokens)

Set reverse DNS

Sets the reverse DNS (PTR) of a server's IPv4 to a hostname, for example mail.example.com. The hostname's A record must already point at the server's IP. Works on Vultr, Linode and Hetzner servers.

Input parameters:

- `hostname` (string, required): Fully qualified hostname, e.g. mail.example.com.
- `id` (string, required): Server id.

### `get_bandwidth` (~50 tokens)

Traffic used this period

Outbound traffic a server used in the current billing period against the traffic its plan includes, in GB. Tracked for Vultr and Hetzner servers.

Input parameters:

- `id` (string, required): Server id.

### `reset_root_password` (~62 tokens)

Set a new root password

Sets a new random root password on a running Linux server, without touching the disk, and returns it once. The password stays in the chat history: log in, add an SSH key and change it. Linux only.

Input parameters:

- `id` (string, required): Server id.

### `get_windows_password` (~49 tokens)

Windows Administrator password

Returns the Administrator password of a Windows server for RDP. Secret: it stays in the chat history, so change it inside Windows after you log in.

Input parameters:

- `id` (string, required): Server id.

### `get_app_password` (~61 tokens)

One-click app password

Returns the admin login and password of a one-click app (for example WordPress or BTCPay) once it has finished installing. Secret: it stays in the chat history, so change it in the app.

Input parameters:

- `id` (string, required): Server id.

### `console_link` (~74 tokens)

Web console link

A fresh web console session for a server (keyboard and screen, works without SSH). Returns a link when the cloud gives one, otherwise the dashboard page where the console opens. The link is a live session: anyone who has it can use the console, so do not share it.

Input parameters:

- `id` (string, required): Server id.

### `reinstall_server` (~149 tokens)

Reinstall the OS

Reinstalls the server's OS in place: the disk is ERASED, the IP stays. Needs confirm: true and the server's exact label; without both it only describes what would happen. Linux: pass ssh_public_key; without it a new random root password is returned once and stays in the chat history. Windows gets a new Administrator password. Runs on Vultr servers.

Input parameters:

- `confirm` (boolean): true reinstalls; without it the tool only describes what would happen.
- `id` (string, required): Server id.
- `label` (string): The server's exact label, as a second check.
- `ssh_public_key` (string): Public key line installed for root after the reinstall.

### `change_ip` (~134 tokens)

Get a new IPv4 address

Moves a server to a new IPv4 address: it is snapshotted and recreated in the same city with the same disk contents, and the old address is released for good. Takes several minutes of downtime. Sets a new random root password, returned once (it stays in the chat history). Needs confirm: true and the server's exact label; without both it only describes what would happen.

Input parameters:

- `confirm` (boolean): true changes the IP; without it the tool only describes what would happen.
- `id` (string, required): Server id.
- `label` (string): The server's exact label, as a second check.

### `list_backups` (~70 tokens)

List backups

Backups (disk snapshots) on the account, or of one server: id, label, status (PENDING, AVAILABLE, ERROR), city, cloud, size, when it was taken and until when it is kept (7 days).

Input parameters:

- `server_id` (string): Only backups of this server.

### `create_backup` (~63 tokens)

Back up a server

Takes a snapshot of a server's whole disk, at no extra charge, kept for 7 days. Works on Vultr cloud servers that are running or stopped. Takes a few minutes; the backup then shows status AVAILABLE.

Input parameters:

- `server_id` (string, required): Server id.

### `restore_backup` (~166 tokens)

Restore a backup to a new server

Creates a NEW server from a backup in the backup's city and cloud, billed per hour like any server. Optionally on a bigger plan. Needs confirm: true; without it returns the price. Refused with the price if it costs more than max_hourly_usd (default $1). Sets a new random root password, returned once (it stays in the chat history).

Input parameters:

- `backup_id` (string, required): Backup id.
- `confirm` (boolean): true creates the server; without it the tool only returns the price.
- `max_hourly_usd` (number): Upper bound for the hourly price, default 1.
- `plan` (string): Plan id or name to restore onto; default the backup's own plan. The disk must be at least as big.

### `delete_backup` (~77 tokens)

Delete backup

Permanently deletes a backup. Needs confirm: true and the backup's exact label; without both it only describes what would be deleted.

Input parameters:

- `backup_id` (string, required)
- `confirm` (boolean): true deletes the backup; without it nothing is deleted.
- `label` (string): The backup's exact label, as a second check.

### `list_load_balancers` (~38 tokens)

List load balancers

Load balancers on the account: id, label, status, city, public IPv4, forwarding rules, backend servers and hourly price.

### `create_load_balancer` (~185 tokens)

Create load balancer

Creates a managed load balancer in a city in front of your servers there, $0.0247/hour ($18.00/month), billed per hour from the balance until it is deleted. Spends money. Backends are ACTIVE vultr or hetzner servers in that city, all on one cloud. Refused with the price if it is above max_hourly_usd (default $1).

Input parameters:

- `algorithm` (string)
- `backend_server_ids` (array): Server ids to balance across; can be added later.
- `forwarding_rules` (array, required): E.g. [{"frontend_protocol":"HTTP","frontend_port":80,"backend_protocol":"HTTP","backend_port":80}].
- `label` (string, required)
- `max_hourly_usd` (number)
- `region` (string, required): City id or name, e.g. "fra".

### `set_load_balancer_backends` (~73 tokens)

Set load balancer backends

Replaces the set of servers behind a load balancer (pass the full new list; an empty list removes all). Servers must be ACTIVE, in the load balancer's city and on its cloud.

Input parameters:

- `backend_server_ids` (array, required)
- `id` (string, required): Load balancer id.

### `delete_load_balancer` (~90 tokens)

Delete load balancer

Deletes a load balancer and its public IP; the servers behind it keep running. Billing for it stops in the hour it is deleted. Needs confirm: true and the exact label; without both it only describes what would be deleted.

Input parameters:

- `confirm` (boolean)
- `id` (string, required)
- `label` (string): The load balancer's exact label, as a second check.

### `billing_history` (~54 tokens)

Balance history

Recent movements of the balance (newest first, up to 50): top-ups, hourly charges, refunds, referral credits, with the balance after each.

Input parameters:

- `limit` (integer): How many entries, default 20.

### `list_invoices` (~46 tokens)

Invoices

Invoices issued for credited crypto top-ups: number, date and amount. The documents open from the Profile page of the dashboard (card payments get a receipt from the card processor instead).

### `list_deposits` (~50 tokens)

Top-ups and their status

Recent top-ups (open and credited, up to 20): amount to pay, amount credited, method, status (PENDING, CREDITED), payment link of open ones and when they expire.

### `request_server_limit` (~85 tokens)

Ask for a higher server limit

Asks for a higher limit on how many servers the account may run at once. Often approved on the spot depending on the account's top-up history; otherwise the team reviews it. One open request at a time.

Input parameters:

- `reason` (string): What the servers are for; helps the review.
- `requested_limit` (integer, required): The new limit (total servers at once).

### `list_payment_methods` (~42 tokens)

Ways to top up

How the balance can be topped up: crypto coins and networks (with their method codes) and the minimum amount for each, and card if offered to this account.

### `top_up` (~177 tokens)

Top up the balance

Creates a payment request for the balance and returns where to pay: a payment link, or an address with the exact amount and network. Nothing is charged until the user pays it. Without sanctions_confirmed the tool creates nothing and returns the sanctions statement that the user confirms. method is a coin and network code (for example usdttrc20) or card. EU customers may see VAT added on top; the result shows the amount to pay.

Input parameters:

- `amount_usd` (number, required): Amount to credit, USD. Minimum 5 for USDT and USDC through our own gateway, 20 for other coins and card.
- `method` (string, required): Coin and network code, e.g. "usdttrc20", or "card".
- `sanctions_confirmed` (boolean): true when the user has confirmed the sanctions statement.

### `support_read` (~63 tokens)

Read support chat

The latest messages of this account's own support chat with NoctHost (oldest first): what the user wrote, replies from the support AI assistant and from the team. Marks replies as read.

Input parameters:

- `limit` (integer): How many recent messages, default 20.

### `support_send` (~95 tokens)

Write to support

Sends a message to NoctHost support in this account's own chat, on the user's behalf; the team sees it marked as sent by an AI assistant. At most 5 messages per hour per key. The support AI assistant often replies at once, and the reply comes back in the result; replies from the team arrive later in the same chat.

Input parameters:

- `text` (string, required): The message, in the user's words where possible.

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/com-nocthost-mcp/mcp#diagnostics

## Score history

- 2026-11-21: 84
- 2026-11-20: 83
- 2026-11-19: 83
- 2026-11-18: 82
- 2026-11-17: 82
- 2026-11-16: 81
- 2026-11-15: 81
- 2026-11-14: 80
- 2026-11-13: 80
- 2026-11-12: 80
- 2026-11-11: 75

## Common questions

### What is the NoctHost MCP server?

NoctHost is an MCP server listed in the public MCP registry as com.nocthost/mcp. Create, list and destroy hourly-billed servers on Vultr, DigitalOcean, Hetzner and Linode. This page covers its hosted endpoint (https://mcp.nocthost.com/mcp).

### Is the NoctHost MCP server safe to use?

NoctHost scores 84 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the NoctHost MCP server expose?

NoctHost exposes 35 tools: get_account, list_options, quote, list_apps, list_servers, and 30 more. Their descriptions and schemas cost roughly 3,683 tokens of context every time the server is loaded.

### Does the NoctHost MCP server require authentication?

Yes. NoctHost asked us for credentials when we connected, so you will need to authorise it in your MCP client before it can do anything.

### Is the NoctHost MCP server still maintained?

NoctHost is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What did VerifyMCP test on the NoctHost MCP server?

VerifyMCP ran 13 checks on NoctHost, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- Remote endpoint: https://mcp.nocthost.com/mcp
- Website: https://nocthost.com/docs/mcp
- Changelog RSS feed: https://verifymcp.io/servers/com-nocthost-mcp/mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/com-nocthost-mcp/mcp/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/com-nocthost-mcp/mcp.json
- HTML version of this page: https://verifymcp.io/servers/com-nocthost-mcp/mcp
