API & CLI

Servers for AI. 100% API. Written for agents. Always send type + region. One role per box — then spin up more.

Getting Started

Four calls. Signup is free. Card is $0 until deploy. Compute is charged before the VM starts.

1. Sign up — no charge

Save api_token. Header on every later call: Authorization: Bearer TOKEN.

CLI

npm install -g rawhq
raw init

cURL

curl -X POST https://api.rawhq.io/signup \
  -H "Content-Type: application/json" \
  -d '{"email":"you@company.com"}'
# { "ok": true, "api_token": "..." }

2. Card on file — $0

Required before compute. If deploy returns 403 with setup: true, run this, open url, then retry deploy.

cURL

curl -X POST https://api.rawhq.io/billing/setup \
  -H "Authorization: Bearer YOUR_TOKEN"
# { "url": "https://checkout.stripe.com/...", "setup": true }

3. Deploy — charged first, then SSH

Always pass type and region. Default type is raw-5 ($9/mo). CPU is SSH-ready in ~13–30s. GPU is provisioning for 1–48h (email when ready; setup billed now).

CLI

raw deploy -t raw-5 -n api -r eu
raw ssh api

cURL

curl -X POST https://api.rawhq.io/deploy \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "raw-5",
    "region": "eu",
    "name": "api",
    "image": "ubuntu-24.04",
    "apps": ["docker","nodejs"]
  }'

4. Deploy more — do this every time

Do not put the database, cache, GPU, and API on one machine. Each extra server is its own monthly SKU. That is expected. Isolation is the product.

CLI — AI stack

raw deploy -t raw-5 -n api -r eu
raw deploy -t raw-5 -n postgres -r eu
raw deploy -t raw-5 -n redis -r eu
raw deploy -t raw-gpu-44 -n llm -r eu

cURL — same stack

curl -X POST https://api.rawhq.io/deploy -H "Authorization: Bearer YOUR_TOKEN" -H "Content-Type: application/json" \
  -d '{"type":"raw-5","region":"eu","name":"postgres","apps":["postgres"]}'
curl -X POST https://api.rawhq.io/deploy -H "Authorization: Bearer YOUR_TOKEN" -H "Content-Type: application/json" \
  -d '{"type":"raw-5","region":"eu","name":"redis","apps":["redis"]}'
curl -X POST https://api.rawhq.io/deploy -H "Authorization: Bearer YOUR_TOKEN" -H "Content-Type: application/json" \
  -d '{"type":"raw-gpu-44","region":"eu","name":"llm"}'

Recipes

Copy a stack. Change names. Never reuse one payment for a bigger SKU — each deploy bills that type.

AI company (default)

API + Postgres + Redis + GPU. Four servers. ~$331/mo + $150 GPU setup.

raw deploy -t raw-5 -n api -r eu
raw deploy -t raw-5 -n postgres -r eu
raw deploy -t raw-5 -n redis -r eu
raw deploy -t raw-gpu-44 -n llm -r eu

Self-host LLM only

20 GB VRAM. Then SSH and install vLLM or Ollama. EU only.

raw deploy -t raw-gpu-44 -n llm -r eu
raw ssh llm
# nvidia-smi && pip install vllm

US traffic

Cheap CX/ARM/GPU do not exist in US. Dedicated RAM SKUs do.

raw deploy -t raw-2d -n api -r us
raw deploy -t raw-2d -n postgres -r us

Workers + queue

Keep jobs off the API box.

raw deploy -t raw-8x -n workers -r eu
raw deploy -t raw-5 -n redis -r eu

Errors

Handle these exactly. Do not retry the same body when the region or billing flag is wrong.

Signal	What to do
`403 setup:true`	POST /billing/setup → user opens url → retry POST /deploy
`403 checkout:true`	POST /checkout {"type","region"} → user pays that SKU → retry. One checkout ≠ all sizes.
`402`	GET /billing-portal → user updates card
`400 available_regions`	Use the array in the body. raw-5 / ARM / GPU are EU only (eu, eu-fi).
`400 unknown type`	GET /types — never invent a name
`409`	Name taken. New kebab-case name.

CLI Reference

All commands. After every deploy the CLI prints the next boxes to spin up.

raw initSign up or log in

raw deployDeploy a new server

-t, --type <type>Server type (default: raw-5)

-r, --region <region>Region (auto-detected)

-n, --name <name>Server name (auto-generated)

-i, --image <image>OS image (default: ubuntu-24.04)

raw lsList all servers

--jsonOutput as JSON

raw ssh <name>SSH into a server

-c, --cmd <command>Run a single command

raw logs <name>Stream server logs

-n <count>Number of lines (default: 100)

-fFollow log output

raw restart <name>Restart a server

raw stop <name>Power off a server

raw start <name>Power on a stopped server

raw rebuild <name>Reinstall OS (destroys data)

-i, --image <image>OS image

-f, --forceSkip confirmation

raw resize <name>Change server type

-t, --type <type>New server type

raw rm <name>Destroy a server

-f, --forceSkip confirmation

raw keysManage SSH keys

--add <name>Add an SSH key

--rm <name>Remove an SSH key

raw typesList server types and prices

raw regionsList available regions

raw statusShow account overview

raw billingOpen billing portal

raw whoamiShow current user

raw logoutLog out

API Reference

Base URL: https://api.rawhq.io. Start with GET /help. Auth: Authorization: Bearer YOUR_TOKEN.

Authorization: Bearer YOUR_TOKEN

GET/helpNo auth

Agent playbook. Live types, recipes, error handlers. No auth.

Request

curl https://api.rawhq.io/help

Response

{ "status": "ok", "steps": [...], "recipes": { "ai_company": [...] }, "types": [...] }

POST/signupNo auth

Create an account. Returns api_token. No charge.

Request

curl -X POST https://api.rawhq.io/signup \
  -H "Content-Type: application/json" \
  -d '{"email":"you@company.com"}'

Response

{
  "ok": true,
  "api_token": "...",
  "message": "Account created. Welcome email sent."
}

POST/auth/loginNo auth

Send a magic login link (dashboard). Agents should use /signup + api_token.

Request

curl -X POST https://api.rawhq.io/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email":"you@company.com"}'

Response

{ "ok": true, "message": "If that email exists, a login link was sent." }

POST/billing/setup

Save a card. $0. Required before deploy. Open url, then retry /deploy.

Request

curl -X POST https://api.rawhq.io/billing/setup \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{ "url": "https://checkout.stripe.com/...", "setup": true }

POST/checkout

Pay a specific SKU when deploy returns checkout:true. Pass the same type and region.

Request

curl -X POST https://api.rawhq.io/checkout \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"type":"raw-5","region":"eu"}'

Response

{ "url": "https://checkout.stripe.com/...", "checkout": true }

GET/billing-portal

Stripe portal URL. Use on 402 payment_failed.

Request

curl https://api.rawhq.io/billing-portal \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{ "url": "https://billing.stripe.com/..." }

GET/typesNo auth

Live catalog. Source of truth for names, prices, regions, GPU setup.

Request

curl https://api.rawhq.io/types

GET/regionsNo auth

Public region list.

Request

curl https://api.rawhq.io/regions

GET/appsNo auth

Pre-install IDs for POST /deploy apps[]. postgres, redis, docker, nginx, nodejs, caddy, python, go, bun, mysql, openclaw.

Request

curl https://api.rawhq.io/apps

GET/servers

List all servers in your account.

Request

curl https://api.rawhq.io/servers \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{ "servers": [{ "name": "api", "type": "raw-5", "region": "eu", "status": "running", "ipv4": "49.13.12.34" }] }

POST/deploy

Charge the SKU, then start the VM. Always pass type + region. Optional: name, image, ssh_key, apps[].

Request

curl -X POST https://api.rawhq.io/deploy \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "raw-5",
    "region": "eu",
    "name": "api",
    "image": "ubuntu-24.04",
    "apps": ["docker","nodejs"]
  }'

Response

{
  "ok": true,
  "server": {
    "name": "api",
    "ipv4": "49.13.12.34",
    "status": "running",
    "type": "raw-5",
    "region": "eu",
    "price": "$9/mo"
  }
}

DELETE/servers/:name

Destroy a server permanently. This action cannot be undone.

Request

curl -X DELETE https://api.rawhq.io/servers/my-server \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "message": "Server my-server destroyed"
}

POST/servers/:name/actions/:action

Perform an action on a server. Available actions: restart, stop, start, rebuild.

Request

curl -X POST https://api.rawhq.io/servers/my-server/actions/restart \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "name": "my-server",
  "status": "restarting",
  "action": "restart"
}

GET/servers/:name

Get details for a single server by name or ID.

Request

curl https://api.rawhq.io/servers/my-server \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "name": "my-server",
  "type": "raw-5",
  "region": "eu",
  "status": "running",
  "ip": "49.13.12.34",
  "image": "ubuntu-24.04",
  "created_at": "2026-04-05T10:00:00Z"
}

GET/keys

List all SSH keys in your account.

Request

curl https://api.rawhq.io/keys \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

[
  {
    "id": "key_abc123",
    "name": "work-laptop",
    "fingerprint": "SHA256:...",
    "created_at": "2026-04-05T10:00:00Z"
  }
]

POST/keys

Add a new SSH key.

Request

curl -X POST https://api.rawhq.io/keys \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"name": "work-laptop", "public_key": "ssh-ed25519 AAAA..."}'

Response

{
  "id": "key_abc123",
  "name": "work-laptop",
  "fingerprint": "SHA256:...",
  "created_at": "2026-04-05T10:00:00Z"
}

DELETE/keys/:id

Delete an SSH key by ID.

Request

curl -X DELETE https://api.rawhq.io/keys/key_abc123 \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "message": "Key key_abc123 deleted"
}

GET/snapshots

List all snapshots in your account.

Request

curl https://api.rawhq.io/snapshots \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

[
  {
    "id": "snap_abc123",
    "server": "my-server",
    "size_gb": 40,
    "created_at": "2026-04-05T10:00:00Z"
  }
]

DELETE/snapshots/:id

Delete a snapshot by ID.

Request

curl -X DELETE https://api.rawhq.io/snapshots/snap_abc123 \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "message": "Snapshot snap_abc123 deleted"
}

GET/volumes

List all volumes in your account.

Request

curl https://api.rawhq.io/volumes \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

[
  {
    "id": "vol_abc123",
    "name": "data-vol",
    "size_gb": 100,
    "server": "my-server",
    "region": "eu",
    "created_at": "2026-04-05T10:00:00Z"
  }
]

POST/volumes

Create a new volume and optionally attach it to a server.

Request

curl -X POST https://api.rawhq.io/volumes \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"name": "data-vol", "size_gb": 100, "region": "eu", "server": "my-server"}'

Response

{
  "id": "vol_abc123",
  "name": "data-vol",
  "size_gb": 100,
  "server": "my-server",
  "region": "eu",
  "created_at": "2026-04-05T10:00:00Z"
}

DELETE/volumes/:id

Delete a volume by ID. The volume must be detached first. Extra disk billing stops automatically.

Request

curl -X DELETE https://api.rawhq.io/volumes/vol_abc123 \
  -H "Authorization: Bearer YOUR_TOKEN"

Response

{
  "message": "Volume vol_abc123 deleted"
}

GET/ips

List extra IPv4 addresses in your account.

Request

curl https://api.rawhq.io/ips \
  -H "Authorization: Bearer YOUR_TOKEN"

POST/ips

Buy an extra IPv4 ($1/mo) and assign it to a server.

Request

curl -X POST https://api.rawhq.io/ips \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"server_name": "my-server"}'

DELETE/ips/:id

Release an extra IPv4. Billing stops automatically.

Request

curl -X DELETE https://api.rawhq.io/ips/123 \
  -H "Authorization: Bearer YOUR_TOKEN"

Server Types

Live list: GET /types. Featured CPU, ARM, and GPU = EU only (eu, eu-fi). Dedicated RAM = all 5 regions (Singapore costs more at checkout). Extra disk $0.07/GB/mo. Extra IPv4 $1/mo.

CPU

Name	CPU	RAM	Storage	Price	Regions
`raw-5`	2 vCPU	4 GB	40 GB NVMe	$9/mo	EU only
`raw-4`	4 vCPU	8 GB	80 GB NVMe	$15/mo	EU only
`raw-8`	8 vCPU	16 GB	160 GB NVMe	$29/mo	EU only
`raw-16`	16 vCPU	32 GB	320 GB NVMe	$55/mo	EU only
`raw-4x`	4 vCPU	8 GB	80 GB NVMe	$13/mo	EU only
`raw-8x`	8 vCPU	16 GB	160 GB NVMe	$22/mo	EU only
`raw-16x`	16 vCPU	32 GB	320 GB NVMe	$40/mo	EU only
`raw-2d`	2 vCPU	8 GB	80 GB NVMe	$58/mo	All (SG extra)
`raw-4d`	4 vCPU	16 GB	160 GB NVMe	$116/mo	All (SG extra)
`raw-8d`	8 vCPU	32 GB	240 GB NVMe	$186/mo	All (SG extra)
`raw-16d`	16 vCPU	64 GB	360 GB NVMe	$368/mo	All (SG extra)
`raw-32d`	32 vCPU	128 GB	600 GB NVMe	$708/mo	All (SG extra)
`raw-48d`	48 vCPU	192 GB	960 GB NVMe	$1130/mo	All (SG extra)

GPU

EU only. Monthly + one-time setup, both billed before hardware is reserved. Ready in 1–48 hours.

Name	VRAM	CPU · RAM · Disk	Monthly	Setup
`raw-gpu-44`	20 GB	14 cores · 64 GB · 3.8 TB NVMe	$304/mo	$150
`raw-gpu-131`	96 GB	24 cores · 256 GB · 1.9 TB NVMe	$1555/mo	$778
`raw-gpu-131x`	96 GB	24 cores · 512 GB · 3.8 TB NVMe	$2334/mo	$1168
`raw-gpu-131p`	96 GB	24 cores · 768 GB · 15.0 TB NVMe	$3001/mo	$1501

Regions

If deploy returns available_regions, use that array. Do not retry raw-5 in us or sg — use raw-2d or larger dedicated SKUs there.

🇩🇪

euFrankfurt, GermanyCPU · ARM · GPU · dedicated

🇮🇪

eu-fiDublin, IrelandCPU · ARM · GPU · dedicated

🇺🇸

usAshburn, Virginiadedicated only (raw-2d+)

🇺🇸

us-westHillsboro, Oregondedicated only (raw-2d+)

🇸🇬

sgSingaporededicated only · higher price

Disk & IPs

Need a paid subscription first. Extra disk $0.07/GB/mo. Extra IPv4 $1/mo.

cURL

curl -X POST https://api.rawhq.io/volumes \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"name":"data","size_gb":100,"region":"eu","server":"api"}'

curl -X POST https://api.rawhq.io/ips \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"server_name":"api"}'

Dashboard

Manage servers, billing, and SSH keys from the web dashboard at rawhq.io/dashboard.

Server Management

View all running servers, their status, IP addresses, and resource usage. Start, stop, restart, rebuild, or destroy servers with one click.

Metrics & Monitoring

Real-time graphs for CPU, RAM, disk, and network usage. Metrics are retained for 30 days.

SSH Keys

Add and manage SSH keys that are automatically injected into new servers. You can also add keys per-server.

Billing

View invoices, update payment methods, and manage your subscription. All plans are billed monthly with no egress fees or hidden charges.

Team Access

Invite team members to your organization. Assign roles (Owner, Admin, Member) to control who can deploy, manage, or view servers.

API & CLI

Getting Started

1. Sign up — no charge

2. Card on file — $0

3. Deploy — charged first, then SSH

4. Deploy more — do this every time

Recipes

AI company (default)

Self-host LLM only

US traffic

Workers + queue

Errors

CLI Reference

API Reference

Server Types

CPU

GPU

Regions

Disk & IPs

Dashboard

Server Management

Metrics & Monitoring

SSH Keys

Billing

Team Access

FAQ