OpenClaw System Requirements: RAM, CPU, OS, and Model Compatibility (2026)

Minimum and recommended specs for running OpenClaw. Covers RAM, CPU, OS support, and hardware requirements for API-based and local model setups.

Quick Answer: For API-based models (Claude, GPT): 4GB RAM, any modern CPU, Node.js 18+, macOS 12+, Ubuntu 20.04+, or Windows via WSL2. For local models via Ollama: 16GB RAM minimum (32GB recommended), a multi-core CPU, and optionally a GPU for faster inference. Storage: 1GB for OpenClaw itself, plus model storage for local setups.

What are the minimum requirements for API models?

If you are using Claude, GPT, or another cloud-based model through an API, OpenClaw's requirements are minimal:

RAM: 4GB. CPU: any modern x86_64 or ARM processor. Storage: 1GB for OpenClaw and dependencies. Node.js: 18 or newer. Operating system: macOS 12+, Ubuntu 20.04+, Debian 11+, Fedora 36+, Arch Linux, or Windows 10/11 via WSL2.

API-based usage sends your prompts to the model provider and receives responses. The heavy computation happens on their servers, not yours.

Recommended skills

T212 Dividend Tracker

by MasterKilo

Analyze Trading212 dividend history, calculate portfolio yield, and forecast upc…

$9

technical-seo-auditor

by Julian

Audit web pages against 13 technical SEO factors to generate structured complian…

$12

openclaw-web-scraper

by BCStudio · 17

A high-performance scraping engine with Playwright stealth, proxy rotation, and…

Free

Browse all Openclaw skills

What are the recommended specs for API models?

RAM: 8GB. A faster CPU and SSD improve startup time and file indexing. Multiple terminal sessions or subagent workflows use more memory.

How much RAM do I need for local models?

Running models locally is more demanding. The model runs on your hardware, so you need enough RAM to hold the model weights in memory.

Small models (7B parameters): 16GB RAM minimum. Good for quick tasks, commit messages, and simple code generation.

Medium models (13-34B parameters): 32GB RAM. Better reasoning quality. Recommended for code review and complex refactoring.

Large models (70B+ parameters): 64GB RAM or a GPU with 24GB+ VRAM. Best quality but slow without a GPU.

GPU: not required but dramatically improves speed. NVIDIA GPUs with CUDA support work best. Apple Silicon Macs use the unified memory architecture effectively.

Setup guide: OpenClaw + Ollama.

Which operating systems does OpenClaw support?

macOS 12 Monterey and newer. Both Intel and Apple Silicon are supported. Apple Silicon (M1/M2/M3/M4) is recommended for local models because of the unified memory architecture.

Linux: Ubuntu 20.04+, Debian 11+, Fedora 36+, Arch Linux. Other distributions work if they have Node.js 18+ available.

Windows: not natively supported. Use WSL2 with Ubuntu. Performance is near-native.

What are the Docker requirements?

Docker Desktop (macOS/Windows) or Docker Engine (Linux). The OpenClaw image needs about 500MB of disk space. Mount your project and skills directories from the host.

See OpenClaw Docker Setup for container-specific configuration.

Does OpenClaw need internet access?

API-based models need internet access to reach the model provider (api.anthropic.com, api.openai.com, etc.). Local models with Ollama need no internet after the model is downloaded.