Full-page rendering, screenshots and JavaScript execution are much heavier than plain HTTP bookmark ingestion.
Users archiving many complex websites.Hardwarevereisten voor Karakeep: RAM, Chrome, zoeken en AI
Lees meer over de hardwarevereisten van Karakeep voor CPU, RAM, de Chrome-crawler, MeiliSearch, SQLite, OCR en lokale AI op ZimaOS.
Karakeep requirements at a glance
Karakeep's current v0.33.0 Docker documentation requires Docker and Docker Compose but does not publish a universal CPU or RAM minimum. The full installation combines the main app with MeiliSearch and browser-assisted crawling; Karakeep stores core data in SQLite and can use local filesystem or S3-compatible asset storage. Optional AI tagging/OCR can call OpenAI or Ollama, but a co-hosted Ollama model must be sized separately.
- CPU
- No numerical official minimum. Browser crawling, OCR, asset preprocessing and search indexing are the main compute bursts.
- RAM
- No numerical host minimum. Current parser settings cap the isolated HTML parsing subprocess at 512 MB by default, but that is not a system RAM requirement.
- Database
- Current architecture uses SQLite for application data and a SQLite-backed job queue.
- Search
- Full installation uses MeiliSearch for full-text indexing.
- Browser crawling
- Full crawling/screenshot/JavaScript execution uses a browser service; without a browser URL Karakeep can fall back to plain HTTP requests without screenshots/JS.
- AI
- AI tagging/OCR can use OpenAI-compatible services or Ollama. Local model RAM/VRAM belongs to the inference backend, not Karakeep itself.
From official requirements to the right setup
Karakeep sizing starts with crawl fidelity, search/index volume and optional AI.
-
Official requirements
Choose full versus minimal installation. The full setup uses MeiliSearch and browser-based crawling, while minimal mode can sacrifice those features to reduce complexity.
-
Confirm your needs
Estimate crawler workload: screenshots, full-page archives, PDFs, video downloads and asset preprocessing all increase CPU, RAM and storage demand.
-
Leave room to grow
Separate AI inference from Karakeep. OpenAI/remote APIs add little local compute; a local Ollama model adds its own RAM/VRAM requirements.
-
Run it on ZimaOS
Install Karakeep from ZimaOS, import a representative bookmark set and measure crawl time, parser memory, MeiliSearch size and asset storage before scaling worker counts.
Check every playback client
- Bookmark count
- Browser crawling enabled
- Screenshots/full-page archives
- MeiliSearch
- SQLite/WAL mode
- OCR
- Asset preprocessing workers
- Remote vs local AI
Official minimum requirements
Karakeep's current v0.33.0 Docker guide lists Docker and Docker Compose as requirements but no numerical CPU/RAM minimum.
Do not turn the 512 MB HTML parser subprocess limit into a host-memory minimum. Full-install resource use comes from several services and optional crawling/AI features.
| Requirement | Official minimum | What this supports |
|---|---|---|
| CPU | No numerical minimum published | Crawler, OCR and indexing workload drive compute. |
| RAM | No numerical minimum published | No host-memory floor is documented. |
| Docker | Required for the documented Docker path | Docker Compose also required. |
| Database | SQLite | Current architecture uses SQLite and a SQLite-backed job queue. |
| HTML parser memory cap | 512 MB default | Per isolated parser subprocess; not total system RAM. |
| GPU | Not required for Karakeep itself | Only a separately hosted local AI model may justify GPU hardware. |
When to upgrade your hardware
Upgrade Karakeep when crawling, indexing, OCR or local AI is measurably slow.
Browser crawling and screenshots push memory/CPU high
Search and asset preprocessing queues grow
Ollama is moved onto the same host
Current config exposes separate search, webhook, asset-preprocessing and rule-engine workers; raising concurrency increases host demand.
Large bookmark collections and automation-heavy users.Local LLM inference changes the hardware class because model RAM/VRAM and inference compute are separate from Karakeep.
Users enabling local AI tagging or LLM OCR.Plan hardware growth with confidence
Scale Karakeep by reducing crawl fidelity, distributing browser/AI work or adding storage only where needed.
Use plain HTTP crawling when screenshots/JS are unnecessary
Without browser URLs, the worker can use plain HTTP requests, skipping screenshotting and JavaScript execution.
Reduces browser CPU/RAM overhead.Tune worker counts carefully
Search indexing, asset preprocessing and rule-engine worker counts are configurable and should rise only with measured headroom.
Concurrency multiplies active CPU/RAM work.Use S3-compatible asset storage
Karakeep supports local filesystem or S3-compatible storage for saved assets.
Useful when screenshots/PDFs/images outgrow local capacity.Keep AI remote unless local inference is a real goal
OpenAI-compatible providers avoid model inference load on the Karakeep host.
Use GPU/local-AI hardware only for a separately justified Ollama workload.Can it run on ZimaOS?
Karakeep is currently available in the ZimaOS App Store under Productivity.
Install Karakeep from ZimaOS
Use the packaged bookmark/read-it-later service with crawling, search and optional AI features.
Open Karakeep in the ZimaOS App StoreUse the current full Docker architecture
Current docs require Docker/Compose and wire persistent storage and service dependencies automatically.
Read Karakeep Docker installationTune crawler/search/AI from current configuration
Current environment variables expose parser memory, worker counts, browser settings, storage and Ollama/OpenAI endpoints.
Read Karakeep configurationChoose Zima hardware for Karakeep
Karakeep can be modest for basic bookmarking but becomes a browser/search/archive/AI pipeline when all features are enabled.
Basic bookmark server or full browser/search/AI archive?
Normal full Karakeep without local LLM
ZimaBoard 2 832 provides useful headroom for Karakeep, MeiliSearch, SQLite and normal crawling.
- Normal Karakeep serverZimaBoard 2 832
- More crawl/index headroomZimaBoard 2 1664
Large archive/storage workload
Use ZimaCube when screenshots/assets and backups drive capacity.
- Large archive serverZimaCube 2 Standard
- Heavier crawl/storage stackZimaCube 2 Pro
No fixed bookmark count or crawl throughput is guaranteed.
| Zima hardware | Best for | Example workload | Core configuration | Recommended boundary | Next step |
|---|---|---|---|---|---|
| ZimaBoard 2 832 | Normal Karakeep server. | MeiliSearch, SQLite and moderate browser crawling. |
|
Heavy screenshot/archive queues can be more demanding. | Get Now |
| ZimaBoard 2 1664 | Larger Karakeep collection and crawler workload. | More browser/index/worker headroom. |
|
Local Ollama models still need separate sizing. | Get Now |
| ZimaCube 2 Standard | Large Karakeep asset/archive collection. | Screenshots, PDFs, assets and backups. |
|
Choose mainly for integrated storage. | Get Now |
| ZimaCube 2 Pro | Heavier crawl/search/storage server. | More CPU, RAM and multi-drive capacity. |
|
Local LLM acceleration is not guaranteed by this tier. | 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
FAQ topics cover RAM, Chrome, MeiliSearch, SQLite, OCR, local AI and storage.
How much RAM does Karakeep need?
Karakeep does not publish a host RAM minimum. Full-install memory includes the app, MeiliSearch and browser crawling.
Is the 512 MB parser limit Karakeep's RAM minimum?
No. It is the default maximum heap for an isolated HTML parsing subprocess.
Does Karakeep require Chrome?
Full browser-assisted crawling uses a browser service, but plain HTTP crawling can operate without screenshots or JavaScript execution.
Does Karakeep require MeiliSearch?
The normal full installation uses MeiliSearch for full-text search; minimal installations can sacrifice features.
Does Karakeep need a GPU?
No. A GPU is relevant only if you separately run a compatible local inference model such as Ollama on the same host.
Can Karakeep store assets on S3?
Yes. Current configuration supports local filesystem and S3-compatible asset storage.
Can ZimaBoard 2 832 run Karakeep?
Yes for a normal full installation without a heavy local LLM workload.
When is ZimaCube 2 useful?
When archived screenshots, PDFs, assets and backups make multi-drive capacity the main constraint.
What sources and further reading informed this Karakeep hardware guide?
Current Karakeep v0.33.0 Docker, architecture and environment-variable docs define the full stack, crawler, parser-memory cap, search workers, storage and AI backends; ZimaOS confirms the current app.
