A standard subscription to a cloud AI assistant costs about $20 a month, $240 a year, just to chat with a model you don’t control while your conversations live on someone else’s servers. The d0a1 packs take the opposite approach: you download the software, run it on your own machine, and the data stays on your hardware. There are four packs, one per use case, and all four install with the same command: ./deploy.sh local. The software is free; the only possible expense comes up if you rent a VPS to keep the service running 24/7.
openwebui: Private AI Chat in Your Browser
This pack installs Open WebUI, an AI chat interface you open in your browser. It does the same things as a commercial chat: ask questions, draft text, translate, or review code. The difference is that the conversation happens on your machine and never passes through a third party’s cloud. It runs on 4 GB of RAM and deploys with ./deploy.sh local; if you want permanent availability, the same script also accepts ./deploy.sh vps. With chat covered, the next pack takes that AI to a website.
wordpress: A Website with Local AI via AI Agent Bridge
This pack installs WordPress and the AI Agent Bridge plugin, which connects the CMS to AI running on your own machine. The typical use case is setting up a blog, a business site, or a personal page and working with AI right from there, without depending on external subscriptions. Because it adds the AI layer, this pack calls for 6 GB of RAM, the highest requirement of the four. Deployment doesn’t change (./deploy.sh local); if the terminal is your thing, the next pack is hermes.
hermes: An AI Agent in the Terminal and on Telegram
This pack installs two pieces: Hermes Agent CLI, an agent you drive from the command line, and its connection to Telegram. The workflow goes like this: you message the agent from the terminal when you’re at your computer and talk to it over Telegram from your phone when you’re not. It needs 4 GB of RAM and you bring it up with ./deploy.sh local; with ./deploy.sh vps, it stays reachable around the clock. If your day-to-day happens in messaging apps, the next pack does away with the terminal entirely.
openclaw: The Assistant Inside WhatsApp, Telegram, Signal, and Discord
The fourth pack installs openclaw, an AI assistant that lives in your messaging apps: WhatsApp, Telegram, Signal, and Discord. The difference from hermes is one of approach: here, chat is the primary point of contact, not a companion to the terminal. It also runs on 4 GB of RAM, and you deploy it with ./deploy.sh local. Choosing among the four comes down to where you spend each day.
Which Pack Fits Each Profile
- openwebui, if you’re looking for a drop-in replacement for paid chat: a browser, 4 GB of RAM, and nothing else.
- hermes, if you work in the terminal and want to reach your agent from your phone.
- openclaw, if you live in WhatsApp or Discord and would rather not switch apps to talk to the AI.
- wordpress, if you manage a website and want to add local AI without signing up for separate services.
All four d0a1 packs share the same deployment command, so trying them one after another costs little more than the installation time.
How to Start Today
Three lines are all it takes:
git clone https://git.d0a1.es/d0a1/packs.git
cd openwebui
./deploy.sh local
With those three commands, you have the private chat up and running on your machine. If you later want the service available 24/7, you run ./deploy.sh vps on a virtual server. To sign up at a discount, d0a1 maintains affiliate links with Vultr and DigitalOcean: the link applies your discount and sends a commission to d0a1.
The Next Step
The barrier to entry is low: a machine with 4 GB of RAM and three commands. Start locally with openwebui to try out the chat in your browser, and save the VPS for when you need permanent availability. When that moment comes, follow the step-by-step deployment guide at d0a1.es, which walks you through the process command by command. The step-by-step deployment guide has the full process.