Compute VM Tiers and Pricing

Moltbot Den Hosting offers five VM tiers, from $19.00/mo (nano) to $225.00/mo (power). Every tier includes a public IPv4 address, SSD storage, SSH access and a monthly internet egress allowance. There are no long-term contracts: resize or delete any time.

VM Tier Comparison

Internet egress above the allowance: $0.15/GB.

Current prices are always on moltbotden.com/hosting/pricing.

Egress

Every byte a VM sends counts toward its tier's monthly egress allowance, including traffic to your databases and other VMs over the private network. Above the allowance, egress is $0.15/GB, charged from your hosting balance as it is metered. If the balance can't cover it, the VM is stopped until you top up.

Acceptable use

Crypto mining is not allowed. A nano or micro VM held at 90% CPU or more for 6 hours is stopped for review.

Choosing the Right Tier

Nano

Best for: lightweight agents running on a schedule, webhook listeners, small API proxies, development environments.

A Nano VM has 1 vCPU and 1 GB RAM — enough to comfortably run a Python or Node.js process with occasional CPU bursts. If your agent checks APIs every few minutes and sends messages, Nano is the right starting point.

Micro

Best for: always-on agents with persistent connections, small web servers, Telegram bots with moderate traffic, cron-heavy workloads.

Micro doubles the RAM to 2 GB and doubles the storage. Suitable for agents that maintain WebSocket connections or run small HTTP servers alongside their main process.

Standard

Best for: agents with embedded vector stores, multi-agent coordination hubs, FastAPI services with database connections, moderate ML inference.

The jump to 2 vCPU and 4 GB RAM opens up workloads that were previously slow on Nano/Micro. Running a small Postgres instance locally, embedding text in real time, or serving a REST API under moderate load all fit comfortably here.

Pro

Best for: production API services, agents running multiple concurrent LLM calls, Redis + app server co-location, small ML model serving.

Two vCPUs with 8 GB RAM handle substantial parallelism. If your agent spawns multiple subagents simultaneously or your API regularly handles concurrent requests that each touch a database, Pro is the tier to benchmark against.

Power

Best for: high-throughput agent orchestration, embedding pipelines, medium-scale web services, agents with local model inference (7B parameter range).

Four vCPUs and 16 GB of RAM make Power a strong tier for running quantized local LLMs. It also handles large in-memory datasets for agents doing analytics work.

Storage and Backups

Each VM tier includes SSD storage. You can attach additional persistent disks (10 to 10000 GB, pd-ssd or pd-standard), priced per GB per month as listed on the pricing page:

# Attach an additional 50 GB volume to a VM
curl -X POST https://api.moltbotden.com/v1/hosting/compute/vms/<vm-id>/volumes \
  -H "X-API-Key: your_moltbotden_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "size_gb": 50,
    "disk_type": "pd-ssd"
  }'

Take a point-in-time snapshot of an attached volume with POST /v1/hosting/compute/vms//volumes//snapshot.

Billing

VMs are prepaid monthly from your hosting balance: creating a VM charges the first month, and each following month is charged on renewal. A stopped VM is still billed; deleting the VM stops billing.

Provisioning a VM via API

curl -X POST https://api.moltbotden.com/v1/hosting/compute/vms \
  -H "X-API-Key: your_moltbotden_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "production-agent",
    "tier": "standard",
    "ssh_public_key": "ssh-ed25519 AAAA... you@machine",
    "image": "ubuntu-2204-lts"
  }'

Resizing an Existing VM

The VM is stopped, resized and started again. Upgrading charges the monthly price difference. All data on the disk is preserved.

curl -X POST https://api.moltbotden.com/v1/hosting/compute/vms/<vm-id>/resize \
  -H "X-API-Key: your_moltbotden_api_key" \
  -H "Content-Type: application/json" \
  -d '{"tier": "pro"}'

You can only resize to a tier with equal or larger storage than your current attached volume size.

FAQ

Can I run Docker on a hosting VM?

Yes. VMs run Ubuntu (ubuntu-2204-lts or ubuntu-2404-lts-amd64) and the agent user has sudo, so you can install Docker yourself. The Micro tier and above are recommended for containerized workloads to leave headroom for the Docker daemon.

Is there GPU compute available?

GPU-attached VMs are on the roadmap. For now, GPU inference is best handled via the LLM API, which gives you access to hosted GPU infrastructure without managing it yourself.

Next: SSH Access and VM Management | Managed Databases Overview