16GB max RAM · 4 bays · 1GbE
AINAS Score 16
No verified price

Traditional NASStationPC
Why Traditional NAS: ARM Rockchip RK3568B2 (4x Cortex-A55) with a 1 TOPS NPU, Mali-G52 iGPU, no GPU path and 4GB LPDDR4 only; AI is limited to photo recognition and search.
Updated
Can the StationPC PocketCloud run local AI?
Not in any useful way. Rockchip RK3568B2 with 4GB of memory at most makes the StationPC PocketCloud a backup and file-sharing box, which is why we class it as a Traditional NAS.
Use it as a baseline when comparing AI-capable alternatives.
The PocketCloud suits someone who wants a small personal cloud for photos and files, not a machine for running language models. Its StationOS software offers AI photo organisation with face and scene recognition, and Smart Search finds photo content with text. Both are vendor confirmed. The hardware is a Rockchip RK3568B2 with 4 cores, a 1 TOPS NPU, 4GB of LPDDR4 and one M.2 NVMe slot for storage. It has no drive bays and uses a 30W supply. Docker is listed, with custom container images and Compose files. A 2-year warranty applies. The AINAS Score is 16, which reflects a system built for light storage and photo tasks rather than local AI. Four configurations are sold, including a 2TB version and a pairing with a Dock.
The 4GB of fixed LPDDR4 rules the PocketCloud out for anyone who wants local language models. Memory cannot be upgraded, the vendor confirms local LLM is not supported, and StationPC staff say local NPU AI is only on the roadmap. Every model listed on this page is an Estimated no, starting with Llama 3.2 3B. Networking is a second point to check. The unit itself has Wi-Fi 6 and Bluetooth 5.3, and the 2.5GbE port comes only with the optional Dock. Buyers who need wired networking from day one should look for a system with built-in 2.5GbE or faster. Anyone who wants local language models needs a different class of machine, with a much higher memory ceiling than 4GB and a GPU path such as a PCIe slot, OCuLink or USB4.
Amazon US
PocketCloud
Ships from the US; delivery and import charges shown at checkout
$269.99
Verified
In stock
Buy on AmazonAmazon US
PocketCloud & Dock
Ships from the US; delivery and import charges shown at checkout
$365.99$409.98
Verified
In stock
Buy on AmazonProduct prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon.co.uk or Amazon.com, as applicable, at the time of purchase will apply to the purchase of this product.
Each capability shows how it was confirmed: Verified by AINAS, Vendor confirmed, Community confirmed or Unverified.
| Capability | Supported | Evidence | Note |
|---|---|---|---|
| Local LLM | No | Vendor confirmed | 4GB RAM ARM; StationPC staff say local NPU AI is only on the roadmap |
| Ollama | No | Unverified | No Ollama support; user request only |
| Open WebUI | No | Unverified | No source; 4GB RAM |
| RAG | No | Unverified | Not offered |
| AI search | Yes | Vendor confirmed | Text-based search of photo content (Smart Search) |
| GPU expansion | No | Vendor confirmed | No PCIe, OCuLink or USB4 |
| Hardware acceleration | Limited | Vendor confirmed | Media Center 4K decoding; transcoding not mentioned |
| AI photos | Yes | Vendor confirmed | AI photo organisation with face and scene recognition |
| AI video | No | Unverified | Not mentioned |
| Surveillance AI | No | Unverified | Not mentioned on StationOS page |
Calculated from the PocketCloud's memory and each model's published size at each quantization. Rows marked Estimated are calculations, not benchmarks. Check any model.
| Model | Best quantization | Runs? | Memory needed | Expected performance | Verification |
|---|---|---|---|---|---|
| Llama 3.2 3BLlama, 3.21B | None fits | No | 11GB at Q4 | Not measured | Estimated |
| Llama 3.1 8BLlama, 8.03B | None fits | No | 14GB at Q4 | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | None fits | No | 15GB at Q4 | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | None fits | No | 18GB at Q4 | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | None fits | No | 20GB at Q4 | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | None fits | No | 19GB at Q4 | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | None fits | No | 25GB at Q4 | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | None fits | No | 29GB at Q4 | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | None fits | No | 31GB at Q4 | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | None fits | No | 60GB at Q4 | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | None fits | No | 89GB at Q4 | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | None fits | No | 65GB at Q4 | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | None fits | No | 96GB at Q4 | Not measured | Estimated |
One verdict per workload, from the PocketCloud's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The PocketCloud is poor for running a local LLM. The local LLM row is no (vendor confirmed), StationPC staff say local NPU AI is only on the roadmap, and Ollama support is not offered, though that row is unconfirmed.
Private RAG is poor on the PocketCloud. RAG is listed as not offered and Open WebUI is marked no, both unconfirmed, and with only 4GB of fixed memory and no local LLM there is nothing to pair retrieval with.
AI search is poor here, despite one working feature. Smart Search, a text-based search of photo content, is vendor confirmed, but it covers photos only and there is no local LLM behind it.
AI photos are rated poor, even though the feature exists. StationOS offers AI photo organisation with face and scene recognition (vendor confirmed), but the 4GB of memory and 1 TOPS NPU limit what the system can do beyond that.
Surveillance is poor on the PocketCloud. The surveillance row is no and unconfirmed, the StationOS page does not mention it, and with 4GB of memory there is no camera recording or detection feature to rely on.
Home lab use is poor. Docker is listed, with custom container images and Compose files, but memory is fixed at 4GB, there are 4 Arm cores, one M.2 slot and no PCIe slot, and virtual machine support is not confirmed.
Business use is poor. There are no drive bays, one M.2 NVMe slot and no ECC listed, and the unit has Wi-Fi 6 only, with 2.5GbE needing the optional Dock. The warranty is 2 years.
Creative AI is poor. GPU expansion is no (vendor confirmed, no PCIe, OCuLink or USB4), hardware acceleration is limited to Media Center 4K decoding with transcoding not mentioned, and local LLM support is no.
4GB of LPDDR4 is all the PocketCloud will ever have; it is not user-upgradeable. Every model you run has to fit inside that figure alongside the operating system.
A graphics card is not an option on the PocketCloud, which has neither a PCIe slot nor OCuLink. Inference runs on the CPU, with Arm Mali-G52-2EE handling media tasks, so keep to small models.
The 1 TOPS NPU helps with photo recognition, transcription and vendor AI features. Most open-source runtimes such as Ollama do not use NAS NPUs yet, so do not count it towards language model speed.
Not well. A useful private RAG setup needs more memory or compute than the PocketCloud offers.
The manufacturer also provides AI search across your files (documented by the manufacturer), which covers simple document questions without setting up your own stack.
For network access it has Not stated. Enough for light use; large imports will take a while.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| PocketCloud | 4GB | None | Diskless, SKU SPC00 (US) / SPC00-EU. |
| PocketCloud & Dock | 4GB | None | With Dock (adds 2.5GbE and second M.2), SKU SPC00-SPC01. |
| PocketCloud 2TB | 4GB | 2TB | 2TB SSD pre-installed, SKU SPC00-2TB. |
| PocketCloud & Dock 2TB each | 4GB | 4TB | 2TB pre-installed in unit and Dock, SKU SPC00-SPC01-2TB. |
Best fit: people who want a private chat assistant or coding model at home without sending data to a cloud service. That is where the PocketCloud scores highest (Local LLM 16/100), followed by RAG (15/100).
Less suited to anyone who wants to run models larger than about 8B parameters and buyers who expect to add a graphics card later.
Other Traditional NAS systems closest in price.
No. Llama 3.2 3B needs 2GB at Q4 and 3.4GB at Q8, and the memory estimate for the PocketCloud is no at both Q4 and Q8 (Estimated). More importantly, nothing on the system can run it. Local LLM support is marked no and vendor confirmed, and StationPC staff say local NPU AI is only on the roadmap.
No. The PocketCloud has 4GB of LPDDR4 and its maximum is also 4GB. No upgrade path is listed, and the memory is not expandable. Every configuration sold, including the 2TB and Dock versions, has the same 4GB, so storage options change but memory does not.
No. The PocketCloud has no PCIe slot, no OCuLink port and no USB4, and GPU expansion is vendor confirmed as not supported. Graphics come from the integrated Arm Mali-G52-2EE, and the only AI hardware is the 1 TOPS NPU in the Rockchip RK3568B2. The USB ports are one USB-A (USB3.0) and a Type-C port, and power arrives through a 30W PD Type-C input. Anyone who wants a graphics card for local AI should look for a system with a PCIe slot, OCuLink or USB4.
Yes. The StationOS page describes custom container images and Compose files, though it does not name any models. Confidence in this is medium. Virtual machine support is not confirmed. With 4GB of fixed memory and 4 Arm cores, containers are limited to light services.
None. There are no SATA bays. Storage is one M.2 NVMe 2280 slot, listed at up to 8TB, and the PCIe generation is not published. The optional Dock adds a second M.2 slot. A 2TB SSD comes pre-installed in one of the configurations.
Four. The PocketCloud is sold diskless, as a 2TB version with the SSD pre-installed, as a PocketCloud & Dock bundle, and as a Dock bundle with 2TB pre-installed in the unit and the Dock. Every configuration has 4GB of memory. The Dock adds 2.5GbE and a second M.2 slot.
No. Ollama support is not yet verified. See the capability table above for the source.
No. Open WebUI support is not yet verified. See the capability table above for the source.
Yes. AI photos support is documented by the manufacturer. See the capability table above for the source.
No. Surveillance AI support is not yet verified. See the capability table above for the source.
Partly. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has Not stated built in. The fastest port is Not stated.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.