Selectel Engineer Showed LLM Agent for Automatic Free Domain Selection

Domain name selection can be transformed from endless manual iteration into a short automated scenario: an LLM generates options, and a script immediately checks their availability and discards occupied ones. This is exactly the pet project described by a Selectel engineer, who put together a simple service for cases when you need to quickly launch a brand, product, or landing page and don't want to spend hours searching for a free address. The project's logic revolves around a quite practical problem.

In popular zones like .com, .ru, .

io, and .org, the space has long been overheated: according to data cited by the author, only in .com there are over 160 million registered domains.

The shortest and most understandable names are taken, and good options are often either occupied or put up for resale. As a result, new projects resort to compromises: adding numbers, extra letters, hyphens, or moving to less familiar zones. The domain market index that the author cites also shows that activity in popular zones has grown approximately threefold over the past three years.

This affects not only brand recognition but also trust: an address should be easy to read, remember, and type without mistakes from a keyboard. The author specifically links the shortage of normal domains to security issues. When a beautiful name is already taken, companies and users get used to forged addresses, mirrors, and questionable variations, and this is fertile ground for phishing and typosquatting.

The article provides examples of domains that differ from the original by a single digit or letter, and also recalls larger cases where attackers deployed hundreds of fake addresses for major events. Against this backdrop, a good domain is no longer purely an aesthetic choice: it's also a way to reduce the risk of mistakes, confusion, and loss of user trust. The service itself is implemented in Python and consists of two steps.

First, the user specifies a project description, desired domain zones, and the number of options. The more precise the initial request, the more meaningful the output: a narrow product description yields more suitable names than abstract categories like "fintech" or "marketplace". Then the LLM through an OpenAI-compatible API generates a list of names according to a strict prompt: short, memorable, without unnecessary characters and only in the specified TLDs.

After this, the second module runs each option through python-whois. If the WHOIS response contains information about the domain, registrar, or creation date, the address is considered occupied; if the service returns an empty response or messages like "not found", the option is marked as available. To avoid waiting too long, the check runs in parallel across ten threads, and the entire cycle usually takes about 10–30 seconds.

On top of this pipeline, the author added a minimalist web interface: a field for the project description, domain zone selection, a counter for the number of results, and a launch button. After submitting the form, the system returns an already-checked list with marks indicating "available" or "occupied". In tests for a query about business coaching in Moscow, the service found, for example, the available option bizmoscow.

org, and for a coffee product store in Saratov offered several more suitable names in the .store, .shop, and .

ru zones. The author himself doesn't attempt to present the project as a finished product: it's more of a working prototype that can be developed through fine-tuning prompts, additional filters, and more careful logic for evaluating name quality. The point of this project is not that AI suddenly solved the domain shortage problem, but that it handles the boring part of the process well.

Here the LLM acts as an idea generator, and WHOIS as a quick reality filter. For startups, solo developers, and small teams, such an approach can save time at launch and reduce the number of pointless manual checks, especially when you don't need the perfect domain for all time, but simply a clear and available address for launch.