Tools We Recommend

Everything you need to run OpenClaw. We've tested all of these. Affiliate links help support this site.

Hardware

Apple Mac mini

Apple Mac mini M6 (32GB)

Best dedicated OpenClaw host. Silent, always-on, low power. 32GB of unified memory runs 27B-class models at Q4.

Check current price
RTX3090

EVGA GeForce RTX 3090 FTW3 Ultra (24GB, Amazon Renewed)

Best value local-LLM GPU. 24 GB VRAM runs Qwen 27B-class at Q4 (~35 tok/sec) — often half a 4090's price used. Used RTX 3090s average $1,626 on eBay, per resaleprices.com, read 12 November 2026. The linked card is an Amazon Renewed unit.

Check current price
RTX4090

ASUS ROG Strix GeForce RTX 4090 (24GB, Renewed)

Same 24 GB as the 3090 but faster (~50 tok/sec on Qwen 27B Q4). The pick if you also game or render. The linked card is a marketplace seller's Renewed unit.

Check current price
96GBPRO

NVIDIA RTX PRO 6000 Blackwell (96GB)

Workstation tier. 96 GB VRAM runs 70B-plus at long context, or several models resident at once on one card.

Check current price

Hosting (Cloud VPS)

DigitalOcean

$200 free credit with referral. Best for beginners. One-click Docker deploys.

Referral link coming soon

AWS Lightsail

$5-20/mo. We have a full video tutorial for this. Great if you're already on AWS.

Watch Tutorial

Hetzner

Cheapest cloud VPS in Europe. Popular with self-hosters. $4-10/mo for OpenClaw.

Link coming soon

AI Models

Local LLM Fit and Speed Estimator

Estimate whether your model fits, where it bottlenecks, and whether to change RAM, VRAM, runtime, quantization, or context length. Open estimator.

Local Model Calculator

Pick your VRAM or Mac memory, a model, a context length from 4K to 256K and an f16, q8_0 or q4_0 KV cache. It adds the cache to the file size and says if the model fits. Open calculator.

Agent Context Budget Checker

Qwen Code, Pi or another agent on Ollama or llama.cpp: does the agent compact before the server runs out of context, and how much room is left for a reply? Open checker.

Tokens Per Second Calculator

Pick a Mac chip or NVIDIA card, a model, a quant and a context. It estimates writing speed from llama.cpp's own benchmark of that device and checks it against the chat, coding and agent floors. Open calculator.

Local AI Payback Calculator

A used RTX 3090, 4090 or 5090, or a Mac: price minus resale after fees, plus electricity, against the API bill it replaces. Three resale cases, and the RunPod price to rent the same card. Open calculator.

Cost per Million Tokens Calculator

Monthly input and output tokens, priced on an API (uncached input, cached input, output) and on your own box (hardware after resale, power, prefill and decode speed). Says which is cheaper and where they break even. Open calculator.

Hybrid Routing Cost Calculator

Keep a frontier model for planning and review, send the rest to a local box or a cheap API. Prices the month three ways and counts failed work done again, review reads, resale and power. Open calculator.

Local Model Team Planner

A router, planner, coder and reviewer on one to three machines. Adds up each machine's models and KV cache, says if they fit together or swap, and gives each machine its endpoint and an OpenClaw config sketch. Open planner.

llama.cpp Command Checker

Paste a llama-server or llama-cli command, or the "error: invalid argument" line. It checks every flag against llama.cpp's own argument table and rewrites removed flags such as --no-mmap and --mlock. Open checker.

Ollama (Free, Local)

Run models on your own hardware with zero API costs. See our benchmarks and recommendations.

Anthropic Claude API

Best tool-calling reliability. ~$6-13/mo for personal use. Pay-as-you-go API keys.

OpenRouter

One API key, access to 100+ models. Good for testing different models without separate accounts.

Guides

Local LLM Fit and Speed Estimator

Model fit, expected bottleneck, runtime, quantization, and context-length triage.

Local Model Calculator

File size plus KV cache at your context length, with the math shown.

Agent Context Budget Checker

Where your agent compacts versus where the server's context runs out.

Tokens Per Second Calculator

Estimated tokens per second for your chip, model and context, and whether it is fast enough.

Local AI Payback Calculator

Does the GPU or Mac pay for itself once you count resale and power?

Cost per Million Tokens Calculator

Local box vs API for your token volume, with input, cache and output priced apart.

Hybrid Routing Cost Calculator

Plan on a frontier model, run the rest locally or on a cheap API: what the split saves.

Local Model Team Planner

Several local models with roles, across one to three machines: what fits at once.

llama.cpp Command Checker

Finds the flag behind "error: invalid argument" and gives the current spelling.

How to Install OpenClaw (2026)

macOS, Linux, Windows. 5-minute install.

Security Checklist: 14 Hardening Steps

Lock down your agent before connecting it to anything.

Fix Every Common Error

Gateway token, blank responses, Docker, macOS, Telegram, Windows.

Some links are affiliate links. We only recommend tools we actually use. Affiliate revenue supports this site.