Tools We Use
The software and platforms behind Southern Claw Labs
Every tool on this page is something we use in our own local AI practice.
No paid placements — just honest recommendations from a working deployment team.
Local-first by default
Where possible, we choose tools that run locally or keep data on Australian infrastructure.
This isn't just ideology — it's practical advice we give to clients and follow ourselves.
Privacy and control matter.
📋 Productivity & Knowledge
How we stay organised and manage knowledge.
Obsidian
Our internal knowledge base and project notes live in Obsidian. Local-first, Markdown-based, and completely private — no cloud sync required. Perfect for a local-AI-first business.
obsidian.md →
Formspree
Our contact form on southernclawlabs.com is powered by Formspree. Takes 5 minutes to set up on a static HTML site, handles spam filtering, and delivers enquiries straight to our inbox. Free tier is plenty for a small business.
formspree.io →
Lemon Squeezy
How we sell the OpenClaw System Play Book and other digital products. Lemon Squeezy handles payments, file delivery, receipts, and tax — all in one platform. No complex setup, AUD supported, and built for small software and digital product businesses.
lemonsqueezy.com →
Screen Studio
Screen recording with automatic zoom, smooth cursor effects, and professional output — perfect for creating AI workflow demos, training videos, and client documentation. We use it to record every tutorial and walkthrough we produce.
screen.studio →🤖 AI & Local Models
The core AI stack we deploy for clients and use ourselves.
Ollama + Ollama Cloud
The easiest way to run large language models locally on a Mac — one command to pull and run any model, private and entirely on your hardware. Ollama Cloud extends this with managed remote access, giving you the same open model ecosystem without needing your own hardware.
ollama.com →
OpenClaw
The AI agent platform that Southern Claw Labs is built around. OpenClaw connects local and cloud AI models to real workflows — document processing, executive assistants, coding agents, and more.
openclaw.ai →
Anthropic Claude
Our cloud AI of choice for tasks that benefit from a frontier model. Used for complex reasoning, long document analysis, and as a coding assistant. Pairs with Ollama for hybrid local/cloud architectures.
claude.ai →
MLX Framework
Apple's Metal-accelerated ML framework — runs AI models natively on the M-series GPU without cloud dependencies. We use MLX for model inference, fine-tuning, and quantization directly on our Mac Mini. Pairs with Ollama for a complete local-first AI stack.
ml-explore.github.io/mlx →
LM Studio
A desktop app for downloading, running, and managing GGUF models on Apple Silicon. Alternative to Ollama with a visual interface — useful for teams exploring different model formats, quantization levels, and fine-tuned models from Hugging Face.
lmstudio.ai →
Whisper (Local Speech-to-Text)
OpenAI's speech recognition model, running entirely on your Mac — no API calls, no cloud, no cost per minute. We use Whisper for transcribing meetings, voice notes, and audio content. Fully local, fully private.
github.com/openai/whisper →
ChatGPT
OpenAI's ChatGPT is part of the AI stack we recommend to clients already familiar with cloud AI tools. We help businesses understand where ChatGPT fits — and when a local or privacy-first model is the better choice.
chatgpt.com →⚡ Infrastructure & Hosting
Where we run and deploy our applications.
Railway
Where our Ollama Model Finder is hosted. Railway makes deploying Node.js apps effortless — push to GitHub and it deploys automatically. Hobby plan starts at $5/month. The easiest deploy platform we've used.
Try Railway →Referral link — you get $20 in free credits
Cloudflare
We migrated southernclawlabs.com DNS to Cloudflare for reliability and speed. Free plan covers everything you need — DNS management, DDoS protection, and fast global propagation. No-brainer for any small business site.
cloudflare.com →
GoDaddy
Our southernclawlabs.com website runs on GoDaddy Economy shared hosting. For a static HTML site it's more than adequate — reliable uptime and simple cPanel file management.
godaddy.com →🛠️ Development
The tools we use to build and ship code.
Visual Studio Code
Our primary code editor for everything — OpenClaw skills, Node.js apps, HTML/CSS, and Python scripts. Extensions like Live Server, Prettier, and Tailwind make front-end work fast. Free, cross-platform, and the most widely supported editor in the AI development ecosystem.
code.visualstudio.com →
GitHub
All our code lives on GitHub. Combined with Railway's auto-deploy, pushing to the main branch is all it takes to ship a new version of the Ollama Model Finder. Version control and deployment in one workflow.
github.com →
Homebrew
The foundation of every Mac development environment. One command to install and manage Python, Node.js, Docker, MLX, and every other tool on this page. If you're deploying local AI on a Mac, Homebrew is where you start.
brew.sh →
Docker
Essential for scaling AI beyond a single Mac. Containerise Ollama AI models, OpenClaw instances, and custom applications for consistent deployment across your team. We use Docker to sandbox agent environments and run isolated services locally.
docker.com →
Apple Mac Mini M4
Our primary development machine. The Mac Mini M4 runs Ollama AI models smoothly, handles all development tasks, and sits silently on the desk. Remarkable performance per dollar for local AI workloads.
apple.com/mac-mini →
Fastmail
Our [email protected] email runs on Fastmail. Privacy-focused, fast, and reliable — a better choice than Google Workspace for a small Australian business that takes data privacy seriously.
fastmail.com →Want us to help you build a local AI stack?
We help Australian businesses select, deploy, and get value from the same tools we use every day.
Contact to discuss an AI project →