Hermes Agent
Paperclip AI
Coolify
Dokploy
WireGuard VPN
Plesk Servers
cPanel Servers
Blockchain NodesDeepSeek 32
$149/ month
- vCores
- 8 dedicated
- RAM
- 32 GB
- Storage
- 240 GB NVMe
- Port
- 500 Mbit/s
Self-hosted LLM
DeepSeek on your own server.
DeepSeek is an open-source family of large language models known for code generation and step-by-step reasoning. Run it on a server with dedicated CPU cores and plenty of RAM, and keep every prompt and answer on your side.
- From $149 / month
- Dedicated AMD EPYC cores
- 32 to 96 GB RAM
- DeepSeek-R1 via Ollama
- token context window
- 128K
- dedicated vCores per server
- 8 to 24
- RAM on the largest size
- 96 GB
- port on the two largest sizes
- 1 Gbit/s
Private DeepSeek Servers
Size the server to your model
CPU inference, no GPU needed. For the 14B DeepSeek-R1 model, start at 32 GB.
DeepSeek 48
$209/ month
- vCores
- 12 dedicated
- RAM
- 48 GB
- Storage
- 360 GB NVMe
- Port
- 750 Mbit/s
DeepSeek 64
$284/ month
- vCores
- 16 dedicated
- RAM
- 64 GB
- Storage
- 480 GB NVMe
- Port
- 1 Gbit/s
DeepSeek 96
$418/ month
- vCores
- 24 dedicated
- RAM
- 96 GB
- Storage
- 720 GB NVMe
- Port
- 1 Gbit/s
Prices in USD, billed monthly unless stated otherwise. Need a different size? Ask for a custom configuration.
What you get
Built in, not bolted on.
Data stays with you
Prompts, documents and model outputs are processed on your server. Nothing goes to an outside AI service.
No per-query billing
One monthly price for the server, whether you send ten prompts a day or ten thousand.
Dedicated CPU cores
Cores and RAM are reserved for your machine, so inference speed does not depend on other customers.
Tune it to your domain
Adjust prompts, add safeguards such as legal disclaimers, or fine-tune a model on your own data.
No vendor lock-in
The models are open source under the MIT license. Switch versions, export them or move to another server when you like.
Isolated setup
Keep the model on a private network with no public endpoint, and log who accesses it.
Use cases
What teams run on it.
-
01
Code review assistant
Let the model read pull requests and suggest improvements without sending source code off-site.
-
02
Support chatbot
Answer customer questions from your own knowledge base while personal data stays on your server.
-
03
Internal search and analysis
Ask plain-language questions about reports, contracts or logs and get answers with context.
-
04
Research sandbox
Give students or analysts an isolated environment for confidential datasets.
Specifications
The details.
| Spec | Value |
|---|---|
| Processor | AMD EPYC 7282, 2.8 GHz |
| CPU cores | Dedicated to your server |
| Storage | NVMe SSD |
| Runtime | Ollama |
| Suggested model | deepseek-r1:14b on 32 GB and up |
| Context window | Up to 128K tokens |
How it works
Three steps to running.
-
01
Choose a size
Pick the RAM your model needs, in Istanbul or Nuremberg. 32 GB fits the 14B DeepSeek-R1 model.
-
02
We install DeepSeek
The server is delivered with Ollama and a DeepSeek-R1 model installed, plus root access.
-
03
Start prompting
Chat over SSH or call the Ollama API from your apps, and pull other model sizes when you need them.
Questions
Private DeepSeek Servers, answered.
Does DeepSeek need a GPU?
No. These servers run DeepSeek on CPU with Ollama. Larger models answer more slowly on CPU, so give them more cores and RAM, or look at our GPU servers.
Which model fits which plan?
The 14B DeepSeek-R1 model is a good starting point on 32 GB. Bigger variants need more memory, so step up to 64 or 96 GB.
Does it need internet access after setup?
No. Once the models are downloaded, DeepSeek runs fully offline. You apply updates when you choose.
Can I use my own OS or DeepSeek version?
Yes. You have root access, so you can install the operating system and DeepSeek version you want and run it through Ollama.
Is it suitable for sensitive data?
Self-hosting keeps data on your own server, which makes privacy rules such as GDPR easier to follow. You control access, encryption and logging.
More app hosting
Often used together.
n8n Hosting
Your own n8n instance on a private server, installed and ready. Unlimited workflows for one flat price.
From $13.20 / month
Explore
Nextcloud Hosting
Your own private cloud for files, calendars and contacts. Nextcloud preinstalled, unlimited users.
From $13.20 / month
Explore
OpenClaw Hosting
A personal AI agent on your own server, reachable from WhatsApp, Telegram, Slack and more. Preinstalled.
From $13.20 / month
ExploreTalk to an engineer
Not sure which size fits?
Tell us what you want to run. An engineer suggests a configuration and sends you a written price.