We’re building name-based filters for domain investors — what are we missing?

[KarmaDomains]

Hi everyone,

We’re the team behind Karma.Domains.

From the beginning, our platform has primarily focused on expired and auction domains for SEO use cases. Most of our current data and filters are related to domain history, backlinks, traffic, content, spam signals, and auction data.

Now we want to build a new group of filters specifically for domain investors. Instead of evaluating only the SEO value of a domain, these filters would focus on the domain name itself.

Before we build them, we’d like to get feedback and criticism from experienced domain investors.

Here’s what we’re currently considering.

Spoiler: filters

Rule-based filters:

• Domain length
• Number of words
• Contains, starts with, ends with, or exactly matches a word
• Position of a word within the domain
• Required and excluded words
• Hyphens and numbers
• Patterns such as LLL, LLLL, NNN, CVCV, and similar structures
• Singular and plural forms
• Prefixes and suffixes
• Selected TLDs

AI-based name classification:

• Language
• Splitting the domain into words
• One-word dictionary domain
• Exact-match domain or keyword phrase
• Acronym or abbreviation
• Brandable or invented name
• Industry or commercial category
• Meaning and possible business use
• Our existing Brand Score

A few simplified examples:

• cheapflights.com → exact-match keyword phrase
• jpps.com.pk → four-letter acronym or abbreviation
• jimly.com → short invented brandable
• pelicano.com.pt → one-word Portuguese dictionary domain
• seyahaton.com → Turkish-derived brandable related to travel

The idea is that investors could combine these name-based filters with auction source, price, TLD, availability, domain history, and other existing data.

We would really appreciate your honest feedback:

1. Which of these filters would you actually use?
2. Which classifications should work as hard filters, and which should only be labels or sorting signals?
3. Which ideas seem too subjective or unreliable?
4. What important name characteristics are we missing?
5. Are there searches you regularly try to perform that existing domain tools do not handle well?

Concrete examples of real searches would be especially useful.

Please be critical. We would rather remove a weak idea now than build something professional domain investors would not trust or use.

Thanks in advance for any advice.

 

[KarmaDomains]

Already ​

Domain name: filters and the breakdown in the report​

The list filters now have a Name tab. It lets you pick domains by the name itself, not by the history of the site.

Domain Name searches the start, the end, and anywhere in the name: a comma means AND, a vertical bar means OR. Next to it are gTLDs/ccTLDs, length, digits, and hyphens.

Letter pattern and Vowel pattern describe the shape of the name. L is a letter, N is a digit, H is a hyphen, X is any other character. In the vowel pattern, V is a, e, i, o, u, y, and C is any other Latin letter.

Name type, Industry, and Name languages are read by an AI model. That is a judgment of the name, not the category of the site, and it can be wrong. Brand Score is filtered in the same block, with a minimum and a maximum.

In the report card the same breakdown is on the Domain name tab.

Structure measures the name without a model: length, the number of digits and hyphens, and the letter and vowel patterns. Each pattern symbol is its own badge; hover it to see what it means.

Classification assembles the name from the words: each word in its own colored block, with a capital letter, and the zone beside them. Below that are the meaning, the type, the languages, and the overall confidence. Under the heading a note says the breakdown was made by AI and may be wrong.