If a representative batch still misses your processing window after reducing unnecessary conversion and tuning concurrency, stronger CPU capacity can help.
Frequent bulk scanning and historical imports.Configuration matérielle requise pour Paperless-ngx : RAM, processeur, stockage, OCR et archives de documents
Planifiez les besoins en RAM, en processeur, en travailleurs OCR, en stockage de documents et en base de données de Paperless-ngx, avec des recommandations pratiques de dimensionnement et de sauvegarde sur ZimaOS pour les serveurs domestiques.
APP
Paperless-ngx hardware requirements at a glance
Paperless-ngx sizing is driven by OCR throughput, worker concurrency and the complete document archive. The reviewed official setup does not publish a numerical whole-host minimum. Size the broker, database, originals, generated archives and backups together.
- RAM
- No numerical official minimum published. OCR processes and optional Tika/Gotenberg need memory alongside the database and broker; do not size only the idle web interface.
- CPU
- No numerical core minimum is published. The v3 migration guide notes an x86-64-v2 baseline for NumPy 2.4 wheels; OCR workers × threads must fit available logical cores.
- Storage
- Budget originals + generated archive files + thumbnails + database/index data + import/export space + independent backups.
- Database / broker
- PostgreSQL is recommended for new installs; SQLite and MariaDB are alternatives. A Redis-compatible broker such as Valkey is required.
- OCR / optional services
- Tika and Gotenberg add Office-document parsing/conversion workloads; they are optional, separate services rather than a GPU requirement.
- Best Zima starting point
- ZimaBoard 2 832 with SSD-backed data is a practical small-archive starting point. Choose 1664 for more services, or ZimaCube 2 Pro for sustained OCR queues.
From official requirements to the right setup
Plan the complete Paperless-ngx workload before selecting a server.
-
Official requirements
Use the current official release and matching Compose files. Prefer PostgreSQL for a new archive and keep the Redis-compatible broker. Prebuilt containers include the application runtime and OCR tools. For v3, set PAPERLESS_SECRET_KEY explicitly and check the migration guide, including the NumPy x86-64-v2 baseline; existing older installations must follow its supported upgrade path.
-
Confirm your needs
Measure pages consumed per day, peak batch size, scan resolution and OCR languages. Time a representative scanned PDF rather than measuring only dashboard response time.
-
Leave room to grow
Persist data and media volumes, consume/export directories, database and broker data as needed. Plan originals and generated documents separately, and keep recoverable backups outside the primary archive.
-
Run it on ZimaOS
Use ZimaOS Install Custom App with the official Compose services. Map container port 8000 to an unused host port, set volume ownership correctly, and limit OCR concurrency so ZimaOS stays responsive.
Check every playback client
- Scanned pages per day
- Peak import batch size
- Scan resolution and OCR languages
- Task workers × threads per worker
- PostgreSQL, SQLite or MariaDB
- Tika/Gotenberg required or omitted
- Originals and generated archive size
- Backup retention and restore test
Official minimum requirements
Checked on October 23, 2026 against official documentation and the project repository. Latest stable release observed: v3.3.4. Numerical requirements below retain their original scope.
No 2 GB or 4 GB host minimum is inferred here. OCR concurrency, database cache and optional converters must fit the actual host. A stored text PDF and a high-resolution scanned PDF can have very different processing costs.
| Requirement | Official minimum | What this supports |
|---|---|---|
| Whole-host RAM | No numerical official minimum published. | Worker count, document complexity and companion services determine practical memory. |
| CPU cores | No numerical official minimum published. | The documented OCR worker/thread controls are throughput settings, not a minimum CPU-core specification. |
| Disk capacity | No numerical official minimum published. | Document count alone is insufficient: scanned page size and archive-generation policy affect capacity. |
| Database | PostgreSQL recommended for new installations | SQLite and MariaDB are also documented deployment choices. |
| Message broker | Redis-compatible broker required | Current upstream Compose examples use Valkey; do not omit the broker. |
| Container / CPU architecture | Official image; check x86-64-v2 compatibility for v3 dependencies | The migration guide documents NumPy 2.4 wheel requirements including SSE4.2. This is an instruction-set floor, not a core-count minimum. |
| Runtime / conversion | Linux; Python 3.11–3.14 for the reviewed bare-metal setup | Use the prebuilt image on ZimaOS. Tika/Gotenberg are optional; build dependencies are not runtime RAM requirements. |
When to upgrade your hardware
Use observed Paperless-ngx workload pressure to decide when an upgrade is warranted.
The OCR queue grows faster than it drains
Workers compete with the database for RAM
Archives and backups approach usable disk capacity
Repeated OOM kills or sustained swapping during imports justify reducing worker concurrency first, then selecting more host RAM.
Multi-app ZimaOS systems or Office-document conversion.Count originals, generated PDF/A files, thumbnails, database/index data and retained backups; leave working room for the next import.
Long-retention document collections.Plan hardware growth with confidence
Address storage, memory, compute and deployment limits separately.
Move active data to SSD
Place the database, index and active processing files on durable SSD storage before assuming CPU is the bottleneck.
SATA/PCIe SSD with ZimaBoard 2.Add host memory headroom
Choose the 16 GB model when OCR workers and companion containers crowd the 8 GB system; the N150 CPU remains unchanged.
ZimaBoard 2 1664.Expand archive capacity
Use multiple drives for a growing archive and retain an independent backup; RAID does not replace document backup.
ZimaCube 2 Standard for storage topology.Increase OCR throughput
Choose a stronger CPU when timed processing stays CPU-bound; increase workers only within CPU and memory capacity.
ZimaCube 2 Pro for sustained OCR and mixed workloads.Can it run on ZimaOS?
No specific public official one-click app page was verified. Zima documentation describes a BigBearCasaOS installation, but a third-party listing is not proof of an official one-click entry. Follow the Custom Install route and current upstream Compose documentation.
Custom install Paperless-ngx in ZimaOS
Deploy the official Paperless image, database and Redis-compatible broker together; add Tika/Gotenberg only when needed. Persist every required path and resolve host port 8000 conflicts.
Custom install Paperless-ngx in the ZimaOS app ↗Use the official Paperless-ngx deployment instructions
Keep the selected stable release, persistent paths and companion services aligned with upstream documentation.
Review Paperless-ngx official installation ↗Control OCR workers and conversion services
Use worker/thread controls to protect CPU and RAM headroom during bulk imports.
Review Paperless-ngx deployment details ↗Which Zima hardware fits Paperless-ngx?
Start with the lowest reasonable host for the actual workload. Product specifications were checked against the current product pages.
Which workload best matches your Paperless-ngx deployment?
Personal archive with occasional scanning
Start with SSD-backed ZimaBoard 2 and conservative OCR concurrency.
- Workload optionZimaBoard 2 832
- Workload optionZimaBoard 2 1664
Large archive or sustained OCR ingestion
Choose capacity for retention, and choose Pro when processing throughput is the constraint.
- Workload optionZimaCube 2 Standard
- Workload optionZimaCube 2 Pro
These are practical workload recommendations, not official app requirements, compatibility certification, measured capacity limits or performance guarantees. Include all co-hosted services and independent backups.
| Zima hardware | Best for | Example workload | Core configuration | Recommended boundary | Next step |
|---|---|---|---|---|---|
| ZimaBoard 2 832 | Personal document archive | Occasional scans, PostgreSQL/broker and low-concurrency OCR. |
|
Use an SSD for active data; the 8 GB configuration is our practical host choice, not an upstream minimum. | Get Now |
| ZimaBoard 2 1664 | More companion services | Paperless plus converters, database and other modest containers. |
|
Extra RAM does not add CPU cores; do not assume more OCR parallelism is always faster. | Get Now |
| ZimaCube 2 Standard | Long-retention archives | Multi-drive original and generated-document retention. |
|
Its 8 GB base RAM is lower than the 16 GB Board; select it for storage, not automatic memory improvement. | Get Now |
| ZimaCube 2 Pro | Sustained bulk OCR | More concurrent processing and a larger mixed document stack. |
|
Processing time still depends on document complexity; no pages-per-minute guarantee is implied. | Get Now |
What the Press Says
Highlights from trusted reviewers worldwide.
“ZimaCube 2: Not just another NAS, tested with 25TB storage, local AI agents, 4K transcoding, and real homelab workflows.”Read full review
“The ZimaBoard 2 is a compact x86 server board that can be turned into a mini NAS, home server, media box, or self-hosting hub.”Read full review
“ZimaCube 2: A modern, high-performance NAS with plenty of room to grow—built for users who want more than basic storage.”Read full review
“Coverage focused on ZimaCube 2's open hardware foundation, no monthly fee, and self-hosting flexibility.”Read full review
Loved by the Community
Stories and reviews from people who build with Zima every day.
★★★★★
Zima Blade Little yet Powerful
Maybe I am not digital natives but I live with PCs since 12 years old in 1984 when IBM PC clone come to my home. Many years have passed and many operating system I've tried. For me Zima blade and CasaOS was a quantum leap for home PC enthusiast and server lab machine to make me stay curious and relevant for this era.
★★★★★
Very good!!
I use ZimaCube Pro as 5th Proxmox cluster node. It runs several VMs and containers, including a VM with GPU passthrough to run a self-hosted LLM. A specific LXC container runs a Samba server for NAS capabilities using four of six RAID 6 SATA HDDs with ZFS.
★★★★★
Great innovation for mini server!
It is very useful and makes a powerful mini server for many purposes, including university and college students in engineering and electronics. Thank you so much for making this server.
★★★★★
Avaliação ZimaBoard 2
Construí um servidor de uso pessoal. O desempenho está muito bom e funciona perfeitamente onde quer que eu esteja. A surpresa é não dependermos de grandes estruturas para termos nosso próprio servidor de dados. Como iniciante, estou gostando bastante do ZimaOS, pois ele é simples e eficiente.
Frequently asked questions
Answers to Paperless-ngx RAM, CPU, storage, Docker, database and ZimaOS sizing questions.
How much RAM does Paperless-ngx need?
No numerical official minimum published. Evaluate peak consumption while OCR, the database, broker and any converters are active. Our 8 GB ZimaBoard starting point includes whole-host headroom and is not an official app minimum.
How many CPU cores should I allocate?
There is no numerical official core minimum in the reviewed setup. Use a timed sample import; keep workers × threads within the logical CPU count and leave capacity for other host services.
Which database and broker are required?
PostgreSQL is recommended for new installations. SQLite and MariaDB are alternatives. A Redis-compatible broker is needed; current upstream Compose templates use Valkey.
How much document storage should I buy?
Measure your scanned originals and generated archive files separately. Add thumbnails, database/index data, temporary import space and backup retention. A 100-page high-resolution scan is not equivalent to 100 small text pages.
Does Paperless-ngx need a GPU or newer CPU instructions?
The standard Tesseract/OCRmyPDF pipeline does not require a dedicated GPU. The v3 migration guide describes an x86-64-v2 instruction-set floor for NumPy 2.4 wheels, including SSE4.2; very old x86 CPUs may fail even without AI features enabled. Optional external AI backends have separate requirements.
Can ZimaBoard 2 run Paperless-ngx?
It is a practical option for a personal archive with SSD-backed data and controlled OCR concurrency. Consider 16 GB when other containers or converters share the machine, while remembering both models use the same N150 CPU.
Why can document imports make the server slow?
OCR, PDF conversion and thumbnail generation are background compute workloads. Lower workers or threads, then check CPU usage, memory pressure and disk latency before buying new hardware.
When is ZimaCube 2 justified?
Choose Standard for multi-drive archive capacity. Choose Pro for measured CPU-bound ingestion and broader service consolidation. A dedicated GPU tier is unnecessary for standard OCR.
What sources informed this Paperless-ngx hardware guide?
Official project requirements, installation/configuration documentation, repository/release information and the current Zima product pages were checked on October 23, 2026. Practical hardware recommendations are editorial sizing guidance rather than upstream minimums.
