You can run DeepSeek-R1 on your own 64-bit Linux VPS without a DeepSeek API key by installing Ollama, downloading a DeepSeek-R1 model variant, and keeping Ollama’s API on the server’s loopback interface. The reliable pattern is: provision enough CPU, memory and persistent storage for the model you choose; install Ollama; pull the exact current model tag; test locally; then add a private network or authenticated TLS reverse proxy only if another machine must connect.
Antihost offers Linux VPS plans with full root access, unlimited traffic on a 1 Gbps port and a 14-day free trial.
What this deployment does
Ollama is the model runner. It downloads and serves model files on your VPS, exposing a local REST API at http://127.0.0.1:11434 by default. DeepSeek-R1 is selected as an Ollama library model, using a tag such as deepseek-r1:7b. Your application can call that local endpoint, so prompts do not have to travel to DeepSeek’s hosted service.
As an Amazon Associate I earn from qualifying purchases.
Free tools Windows power users keep installed
One-click scans. No signup required.
This is different from using DeepSeek’s hosted API. The hosted service is an authenticated, OpenAI-compatible endpoint at https://api.deepseek.com; it requires an API key and follows the model names and service rules published by DeepSeek. A VPS Ollama installation is self-managed: you provide the compute, disk, security controls, upgrades, backups and monitoring.
Requirements and sizing before you create the VPS
Start with a supported 64-bit Linux distribution, SSH access, a user with sudo privileges, and a persistent disk. Ubuntu is a common choice, but the installer is intended for Linux generally. Install only after you have selected a model variant and understand the storage and memory consequences.
#1 Best Overall
- BLUE LIGHT BLOCKING GLASSES: Our blue light glasses for women boast UV400 protection, blocking harmful blue rays from computer screens. With these bluelight glasses women, you can enjoy gaming, reading, or watching TV with no worries of eye strain.
- EYE RELIEF: Our blue light glasses are engineered to reduce blurry vision or strain caused by prolonged screen time. These womens blue light glasses aid in better sleep by preventing blue light from disrupting your natural sleep cycle, helping you fall asleep faster and enjoy a more restful night's sleep, making them the perfect christmas present.
- COMFORTABLE: Crafted from durable material, our trendy blue light glasses for woman or blue light glasses men ensure long-lasting comfort. The high-transmittance lenses accessories, gamer gifts for men, help keep your eyes relaxed. Perfect stocking stuffers, gifts for gamers men, christmas gifts for men, halloween costumes.
- LIGHTWEIGHT: These clear blue light glasses feature a lightweight frame, meaning they’ll stay comfy for long-term wear. Perfect as blue blockers glasses for women or mens blue light glasses, these non-prescription fake glasses for men can be worn throughout the day. The blue light blocking glasses women feature polycarbonate HD lenses, which are anti-reflective and restore true color.
- WORRY-FREE PURCHASE: If you experience any issues, our support team is ready to assist you quickly. These are the perfect costume glasses for cosplay dressing up.
There is no universal RAM or VRAM minimum
The official Ollama and DeepSeek material does not publish one RAM requirement, VRAM requirement, tokens-per-second figure or monthly VPS price that applies to every DeepSeek-R1 variant. Actual needs vary with parameter size, quantization, context length, concurrent requests, operating-system overhead and whether inference runs on CPU or GPU. Treat a provider’s advertised memory as a starting point, not a guarantee that a particular tag will load or perform acceptably.
| Decision | What changes | How to size it |
|---|---|---|
| Model variant and tag | Download size, working memory and response quality | Check the current DeepSeek-R1 entry in Ollama’s model library immediately before deployment; tags and variants can change. |
| Quantization | Memory footprint, speed and sometimes output quality | Use the quantized variant that fits your available memory while leaving headroom for the OS and context. |
| Context length | Additional memory per request | Budget for your application’s longest prompts, not only a short hello-world test. |
| Concurrency | More simultaneous requests increase CPU, RAM and GPU pressure | Begin with one worker and measure before allowing parallel traffic. |
| GPU access | Potentially lower latency, but adds driver and instance constraints | Choose a VPS with supported NVIDIA hardware or AMD ROCm support only when your selected model and workload justify it. |
| Disk | Model files persist between restarts and can consume substantial space | Use persistent storage sized for every model you intend to keep, plus temporary download and operating-system headroom. |
Do not choose a tiny instance on the assumption that a “7B” label guarantees a fixed memory requirement. Confirm the model’s current file size and leave room for the context window and other services. If the provider offers GPU instances, verify that the advertised GPU is actually passed through to the guest and that its driver stack is supported.
1. Provision and prepare the VPS
- Create a 64-bit Linux VPS in the region required by your users or compliance policy. Attach persistent storage rather than relying on an ephemeral boot disk.
- Record the private and public network addresses. Your application should use a private address, VPN or internal network whenever possible.
- Log in over SSH with a non-root account and confirm that
sudoworks. - Apply the distribution’s security updates, set the correct time zone and enable a host firewall. Keep SSH restricted to your administration source addresses if your provider supports that control.
- Check free memory and disk before downloading a model:
free -h
df -hDo not open TCP port 11434 at this stage. Ollama is designed to listen on loopback by default, and your first test should occur on the VPS itself.
2. Install Ollama on Linux
Native installation
Ollama’s official Linux installer is:
curl -fsSL https://ollama.com/install.sh | shThe command downloads and executes the installer supplied at https://ollama.com/install.sh. Review your organization’s software-installation policy before piping a remote script to a shell; an alternative is to download it first, inspect it, and then execute it according to your policy.
After installation, check whether the package created a systemd service:
sudo systemctl status ollamaOn a systemd host, start it at boot with:
sudo systemctl enable --now ollamaIf your installation does not create a service, run the server in a managed process instead:
Recommended Free Tools
Rank #2
- WINDEX GAMING WIPES: effortlessly clean all your gaming gear and gaming electronics in one easy step with our new Windex Gaming Wipes
- AMMONIA FREE: our ammonia-free formula is safe to use on all of your favorite gaming equipment like consoles, headsets, controllers, screens, keyboards, and more
- PRE-MOISTENED WIPES: each wipe instantly removes dust, dirt, and stubborn smudges and fingerprints leaving a streak-free shine
- ANTI-STATIC: great for cleaning and wiping down your gaming gear and electronics leaving behind no visible residue
- COMPACT RESEALABLE PACKAGE: store your pack of Windex Gaming Wipes anywhere in your home for whenever a quick and easy wipedown is needed
ollama serveKeep that process supervised by your init system or process manager in production. Running it only inside an SSH session means it will stop when the session ends unless you deliberately use a supervisor.
Docker installation
Ollama also publishes the official Docker image ollama/ollama. A container deployment must persist the model directory and expose the API only to the interfaces you intend to use. For NVIDIA GPU acceleration on Linux, Ollama’s FAQ states that the NVIDIA Container Toolkit is required. Containerizing Ollama does not remove the need for firewalling, authentication or TLS at the network boundary.
3. Pull and start a DeepSeek-R1 model
Use the exact tag displayed on the current Ollama DeepSeek-R1 library page. The following example uses deepseek-r1:7b; it is an example tag, not a promise that the same tag or variant will remain available:
ollama pull deepseek-r1:7b
ollama run deepseek-r1:7bollama pull downloads the model files to the server. ollama run starts an interactive session and will pull the model first if it is not already present. Replace the tag in every later command with the tag you actually selected. Model names, tags and available variants can change, so the library listing is the authority at deployment time.
To see what is installed locally, use:
ollama listLeave the interactive session with the normal exit command for your shell, then test the HTTP API independently. Keeping the model tag in an environment variable can reduce copy-and-paste mistakes in scripts:
MODEL='deepseek-r1:7b'
ollama run "$MODEL"4. Test the REST API on the VPS
Ollama’s documented chat pattern sends JSON to the local endpoint. Run this on the VPS:
Rank #3
- Ergonomic & Adjustable Design: Fight "text neck" and improve your posture. Designed for comfort, it offers an adjustable height (7.1in-8.5in) (1.4in adjustable range) and angle (5°-85°) Please note: cell phone stand is only 9 inches tall—the length of a standard pencil
- Without blocking subtitles Perfect for office desk Tiktok Face Time Zoom video—this is the ultimate kitchen desk essential for phones! Perfect viewing experience while reducing neck back strain, A must-have for gamers streamers desk, and anyone who values both style and support!
- Sleek & Stable Design fits Otterbox Case: Say goodbye to bent cables, tangles, and tip-overs! LISEN’s stylish adjustable desk phone holder features a built-in charging hole and smart cable management, keeping your phone secure and your workspace clutter-free cool gadgets for men
- ANTI-SLIP DESIGN free hands: The under 10.00 items easy gadgets back and the bottom of this kitchen essentials for women are fully covered by anti-skid silicone, which LISEN stand up desk decor can provide maximum protection device from any scratches and slides
- Your Perfect Viewing Companion:If your product arrives damaged or breaks within 365 days, reach out to LISEN for a quick and hassle-free solution. Struggling with assembly, like tightening the screw? We’ve got your back! Your experience matters to us. (Product including: phone Holder stand*1, screw*1, hexagon screwdriver*1, specification*1)
curl http://127.0.0.1:11434/api/chat
-H 'Content-Type: application/json'
-d '{"model":"deepseek-r1:7b","messages":[{"role":"user","content":"Say hello"}],"stream":false}'A successful response is JSON containing the assistant message and generation metadata. If you receive “connection refused,” the service is not listening; check sudo systemctl status ollama or start ollama serve. If the response says the model is missing, compare the JSON model value with ollama list and pull that exact tag.
Keep stream set to false when a client expects one complete JSON response. Streaming can be useful for an interactive UI, but it changes response handling and should be enabled deliberately in the client.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
5. Connect an application without exposing Ollama
The safest first architecture is to run your application on the same VPS and point it to http://127.0.0.1:11434. No public firewall rule is needed for Ollama, and packets never leave the host. If the application runs on a different machine, use a private network, site-to-site VPN or an authenticated reverse proxy.
Reverse-proxy pattern
Place Nginx, Caddy or an equivalent proxy in front of Ollama. The proxy should terminate TLS, authenticate callers and forward only approved requests to http://127.0.0.1:11434. A minimal Nginx location uses:
location / {
proxy_pass http://127.0.0.1:11434;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
}This snippet is only the forwarding portion. Add a real certificate, access control and request-size/time-out policy appropriate to your application before making the proxy reachable. Do not treat an obscure URL as authentication. If your clients can use a VPN or private network, prefer that path and keep the proxy off the public internet.
Changing Ollama’s bind address
Ollama documents the OLLAMA_HOST setting for changing its listening address. The default is loopback; the vendor states, “Ollama binds 127.0.0.1 port 11434 by default.” Only change the bind address when a specific network design requires it. Binding to all interfaces can make the API reachable from every network attached to the VPS, so pair the change with firewall rules and an authenticated proxy.
Rank #4
- Easy Installation: Our gaming lights have been designed to enhance your entertainment experience. Simply position the led light bar next to your TV or game console to create a game room environment with ease. They come with USB power and a 60 inch cord for direct mounting and mobility
- Audio Syncing & Extensive Mode: The built-in high-sensitivity music sensor allows you to sync the TV led lights to your favorite music. Lights react and move to the beat of music, film and game audio, creating an incredible immersive sound and light experience – and all without the need for an additional hub. You can choose from 64+ patterns, including strobe, illusion, jump, energy, gradient, running, chasing, fluttering mode etc. Each mode can be applied to different scenes
- 64+ Chasing IC+RGB & W Effect: The independent control design of IC+RGB and white all-in-one multifunctional rgbw led bar lights monitor backlight allows you to easily switch the led lights for desk between IC+RGB and white to enjoy a variety of lighting effects, you can choose white lighting mode with high brightness for better illumination when working, etc. Choose IC+RGB lighting mode when you need to create an entertaining atmosphere for the light show, 16 million colours of Color changing rgb light bar let you dance to the beat and create a dynamic party atmosphere
- Dynamic and Dimmable Lights: 1%-full brightness free brightness adjustment gives you the flexibility to adjust dimmable led light bar for room decorations or a variety of occasions
- USB Powered: USB light bars for room can be powered by powernank, usb charger, TV, Monitor, laptop PC mainframe, making behind TV led lighting bar easy to install in any location.TV led lights are powered by the TV usb interface / PC mainframe, can be turned on/off synchronously with TV / PC
For a systemd installation, create a service override with sudo systemctl edit ollama:
[Service]
Environment="OLLAMA_HOST=127.0.0.1:11434"After saving, reload and restart:
sudo systemctl daemon-reload
sudo systemctl restart ollamaIf a private network requires a non-loopback bind, substitute the approved private address and verify from another private host before changing any public firewall rule. Never publish port 11434 directly to the internet.
6. Firewall and security checklist
- Allow SSH only from administration addresses where possible.
- Allow the TLS proxy or VPN port, normally 443 for HTTPS, only from intended clients.
- Do not allow inbound TCP 11434 from the public internet.
- Require authentication at the proxy or private-network boundary; Ollama’s local listener should not be mistaken for an internet-facing access-control system.
- Terminate TLS before traffic crosses an untrusted network.
- Patch the operating system, Ollama, proxy and GPU drivers on a planned schedule.
- Monitor disk usage, memory pressure, restart events and authentication failures.
- Back up application configuration and any custom model definitions. Re-downloadable model files still need enough persistent storage after a rebuild.
A simple UFW policy might allow SSH and HTTPS while leaving 11434 closed:
sudo ufw allow OpenSSH
sudo ufw allow 443/tcp
sudo ufw enable
sudo ufw status verboseAdapt these rules to your provider’s firewall and your actual SSH port. Applying a firewall remotely can disconnect you if the SSH rule is wrong, so verify the administrative path first.
7. Keep model files on persistent storage
On Linux, Ollama documents /usr/share/ollama/.ollama/models as the default model location. If the VPS boot disk is small or disposable, attach a persistent volume and relocate the model directory with OLLAMA_MODELS. The Ollama service account must be able to read and write the destination.
- Format and mount the persistent volume using your provider’s documented filesystem procedure.
- Create a directory for Ollama’s models and assign ownership to the
ollamauser:
sudo mkdir -p /srv/ollama-models
sudo chown -R ollama:ollama /srv/ollama-models- Create a systemd override with
sudo systemctl edit ollama:
[Service]
Environment="OLLAMA_MODELS=/srv/ollama-models"- Reload the unit, restart Ollama and pull the model again if it is not present in the new location:
sudo systemctl daemon-reload
sudo systemctl restart ollama
ollama listConfirm the volume mounts automatically at boot before relying on it. A path that exists only until the next restart can cause Ollama to appear empty and trigger an unnecessary re-download.
Best Value
- Compatible with Wide Screens - To ensure compatibility with the dual monitor mount, your each monitor must meet three conditions at the same time: First, computer screens size range: 13 to 32 inches. Second, screen weight range: 4.4 to 19.8 lbs. Third, the back of the monitor screen must have VESA mounting holes with a pitch of 75x75mm or 100x100mm.
- Regarding the compatibility with desks - Your desk must meet three conditions at the same time: First, desk material: Only wooden desks are recommended, plastic or glass desks cannot be used. Second, desk thickness range: 0.59" - 3.54". Third, the bottom of the desk should not have any cross beams or panels, as this will interfere with installation. We recommend carefully checking that your desk and monitors meets all above conditions before purchasing.
- Dual C-Clamp Hold - Worried your dual monitors might wobble or slip? Our upgraded base uses a larger platform plus a dual C-clamp structure to lock the dual monitor arm firmly to your desk. Each arm safely keeps your screens steady while you type, click and game—no shaking, no sliding, just a clean and secure setup you can trust every day. It also provides Grommet Mounting installation choice, both options ensure stable and secure fixation for your 0.59" - 3.54" desk.
- Full-Motion Adjustment For Comfortable View - Pull the screen closer when you’re deep in a spreadsheet, push it back to watch videos, or rotate to portrait for coding — moving everything smoothly with just one hand. The monitor stand offers +85°/-50° tilt, ±90° swivel and 360° rotation. Raise your monitor up to 15.75″ to support a healthy sitting posture. Whether you’re working from home, gaming through the night, or switching between video calls and documents, getting the screens to your natural line of sight helps relieve neck, shoulder and back strain so you can stay focused longer with less fatigue.
- Keep Your Desk Organized: By lifting both screens off the desktop, this dual monitor stand opens up valuable space for your keyboard, notebook, docking station or a simple, clutter-free work area. Built-in cable management guides wires along the arms, keeping cords out of sight and out of the way. Enjoy a tidy, modern workstation that looks as good as it feels to use.
8. Force local-only operation
If your privacy policy requires Ollama to avoid cloud models and web search, set OLLAMA_NO_CLOUD=1 and restart the service. For a systemd installation:
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutesudo systemctl edit ollama[Service]
Environment="OLLAMA_NO_CLOUD=1"sudo systemctl daemon-reload
sudo systemctl restart ollamaOllama also documents a disable_ollama_cloud setting in ~/.ollama/server.json. Use one deliberate configuration method, document it, and verify after upgrades that the setting remains active. Disabling cloud features does not secure a publicly exposed port: network isolation, authentication and patching remain your responsibility.
9. GPU acceleration and CPU fallback
Ollama documents NVIDIA GPU support and AMD GPU support through ROCm on Linux. A GPU-enabled VPS can reduce latency, but the correct instance depends on the model’s quantization, context length and concurrency. Driver compatibility, device pass-through and provider quotas can matter as much as nominal GPU memory.
- Choose CPU execution when the workload is light, latency is unimportant or a supported GPU is unavailable.
- Choose a GPU instance when sustained interactive traffic justifies its cost and the selected model fits the available device memory.
- For Docker on Linux with NVIDIA hardware, install the NVIDIA Container Toolkit as required by Ollama’s FAQ.
- Use Ollama’s documented environment controls for selecting a GPU or forcing CPU execution; verify the result with a real request rather than assuming the device is being used.
Measure your own workload after deployment. A short prompt with one user does not predict performance for long contexts, multiple concurrent users or tool-heavy application traffic.
10. Operating the service after installation
Service health
sudo systemctl status ollama
sudo journalctl -u ollama --since '15 minutes ago'Use the journal to distinguish a failed process from a model-loading error, permission problem or out-of-memory kill. Keep an eye on free -h and df -h during first pulls and representative requests.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Updates and model changes
Ollama releases and model-library tags can change. Record the Ollama version, model tag, quantization, context settings and operating-system image used by each environment. Pull a new tag into a staging directory or staging VPS, run representative prompts, then promote it rather than replacing a working production model blindly.
Backups and rebuilds
Back up service overrides, proxy configuration, firewall policy and application secrets. Model blobs can often be downloaded again, but a rebuild still requires time, bandwidth and enough disk. A persistent volume protects against a boot-disk replacement; it does not protect against deletion, corruption or provider failure, so follow your normal backup policy.
Common failures and fixes
| Symptom | Likely cause | Action |
|---|---|---|
curl: connection refused on 127.0.0.1:11434 |
Ollama is stopped or listening elsewhere | Run sudo systemctl status ollama; start the service or ollama serve; inspect the journal and confirm the bind setting. |
| Model not found | The tag in the request differs from the pulled tag | Run ollama list, copy the exact tag and pull it before testing again. |
| Download stops with a disk error | Insufficient space or a non-persistent volume | Check df -h, expand or mount persistent storage, set OLLAMA_MODELS and verify ownership. |
| Process is killed during generation | Memory pressure from model, context or concurrency | Reduce context or concurrency, select a smaller/ more heavily quantized variant, add RAM or use a suitable GPU instance. |
| Remote client cannot connect | Loopback-only bind, firewall, proxy or private-network route | Keep loopback when possible; otherwise verify the private route, proxy upstream, TLS certificate, authentication and firewall without opening 11434 publicly. |
| GPU is idle | Unsupported driver, missing pass-through, wrong container runtime or model too large for the device | Check provider GPU support, Linux driver/ROCm or NVIDIA Container Toolkit requirements, then verify with a representative request and system telemetry. |
| Cloud requests appear despite a local deployment | Cloud features were not explicitly disabled | Set OLLAMA_NO_CLOUD=1 or disable_ollama_cloud, restart Ollama and document the active setting. |
Local Ollama on a VPS versus DeepSeek’s hosted API
| Axis | Ollama on your VPS | DeepSeek hosted API |
|---|---|---|
| Data path | Prompts can remain on infrastructure you control when cloud features are disabled; the operator secures the host and network. | Requests go to DeepSeek’s hosted endpoint and require an API key. |
| Billing model | You pay for the VPS, storage, bandwidth and optional GPU; the cited official material does not establish a universal monthly price. | Usage is metered by the provider’s current hosted pricing; no price should be assumed from a local deployment. |
| Maintenance | You maintain Linux, Ollama, model files, drivers, proxy, backups and monitoring. | The provider operates the inference service; you manage credentials, client code and usage controls. |
| Latency | Same-host calls avoid an internet round trip, but generation speed depends on your VPS hardware and workload. | Network distance and provider queueing affect latency; hardware is not yours to tune. |
| Version control | You choose when to pull a model tag and can pin a tested local setup, while tags themselves may change. | You use the hosted model names and behavior exposed by the API at the time of the request. |
| Scaling | Scale by resizing the VPS, adding GPU capacity or deploying additional Ollama workers behind your own routing. | Scale through the provider’s service limits and account controls. |
| GPU requirement | Optional; CPU works where the model fits, while NVIDIA or AMD acceleration can reduce latency. | No customer GPU is required. |
| API authentication | A local listener is not an internet authentication layer; add proxy, VPN and access controls for remote clients. | Authentication is part of the hosted API connection through an API key. |
Which path should you choose?
- Choose Ollama on a VPS when prompts must stay within infrastructure you administer, you need control over model files and versions, or your application benefits from a same-host endpoint.
- Choose the hosted API when you do not want to operate model-serving hardware, need provider-managed scaling, or cannot maintain drivers, storage and security controls.
- Use a hybrid design when sensitive prompts stay local but non-sensitive workloads use the hosted endpoint. Keep the two paths explicit in code so an outage or configuration change cannot silently route private data to the wrong service.
For a first deployment, keep Ollama bound to 127.0.0.1:11434, pull one verified DeepSeek-R1 tag, test the local REST request, and only then add a private network or authenticated TLS proxy. That sequence gives you a working baseline and prevents an unfinished model server from becoming an unauthenticated public service.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →
