Both approaches have a place. This page explains the difference fairly, shows a comparison, and helps you decide which fits your work. We do not quote EstiBot's prices, because they change and we have not verified them today. Check their site for current pricing.
What EstiBot does well
EstiBot is built for speed at volume. You can appraise large lists of names automatically, and its API lets other tools request appraisals. For an investor screening thousands of drops, or a platform that needs a number for every listing, a fast automated estimate is useful. Many investors use an automated estimate as one signal among several.
Where a model estimate leaves you
An automated model gives you a number. What it usually does not give you is the evidence behind that number in a form you can check. When a model estimate is close to what similar names sell for, that is fine. When it is far off, you have no easy way to see why.
That matters most in two moments. Before you bid at an expired auction, an estimate that is too high can push you into overpaying. When you talk to a buyer or a seller, a number without sales behind it is hard to defend. A broker cannot hand a client a model output and say "trust it".
How our estimate is built
SoldComparables starts from 128,867 recorded domain sales between 2021-05-28 and 2026-05-28, taken from public NamePros "Daily Domain Sales" reports, which republish NameBio top sales charts. Every recorded sale is at least 1,000 USD, the floor of those reports. Each one shows its venue and links to the public report thread it came from. We do not invent sales.
For your name, the search finds semantically similar sales. An embedding language model has precomputed which words are close in meaning, so "cron" is near "scheduler", "task" and "job". Each comp gets a relevance score from 0 to 100 and up to three reasons, such as "Same .ai extension", "Shares the word cron" or "Same length".
The estimate then gives:
- a low, likely and high value in USD
- a wholesale figure and a retail figure
- a confidence level, High, Medium or Low
- a plain English rationale that cites the comps by name, price and date
- what would raise or lower the number
When fewer than three good comps exist, it says so, widens the range and lowers confidence. When no comparable recorded sale exists above 1,000 USD, it says names like this usually trade below that and does not produce a figure. It is a data-backed estimate, not a certified appraisal and not a guarantee of sale price.
Why inspectable comps change the conversation
Developer-term .ai names show how much dates and context matter. stack.ai sold for 4,449 USD on Whois.ai (2021-07-06) and for 72,999 USD on Dynadot (2024-06-03). debug.ai sold for 7,260 USD on Whois.ai (2023-05-08) and for 15,000 USD on Whois.ai (2023-06-05). A single number hides that history. A list of comps with dates shows it, and you can decide which ones apply to your name.
Comparison table
| EstiBot | SoldComparables | |
|---|---|---|
| How the estimate is made | Automated model | Built from cited sold comparables |
| Sales behind the number | Check their site | Every comp listed with price, date, venue and source link |
| Relevance of each comp | Not applicable | Score 0 to 100 with up to 3 reasons |
| Wholesale and retail | Check their site | Separate wholesale and retail figures |
| Confidence and gaps | Check their site | Confidence level; says so when comps are thin or missing |
| Bulk | Yes | Up to 200 names on Investor, up to 1,000 on Broker |
| API | Yes | REST API with 5,000 calls per month on Broker |
| Reports | Check their site | PDF and CSV; white label PDF on Broker |
| Pricing | Check their site for current pricing | From 49 USD per month, 24 USD per month billed yearly |
When to use which
An automated model fits a first pass over a very large list where you only want to sort names into rough buckets, and where reading evidence for each name is not practical.
SoldComparables fits decisions where money moves: a bid ceiling at an expired auction, a buy now price on a name you expect to sell, a renewal you are unsure about, or a price you need to justify to someone else. See domain investing tools for portfolio work and domain broker tools for client reports.
Many investors will use both: a model to narrow a list, then scored comps to decide on the names that survive. Our page on why automated domain appraisals overvalue names goes deeper into the gap between model output and recorded sales.
Bulk and API on our side
If you came to EstiBot for bulk and API, here is what we offer in numbers.
| Plan | Monthly | Billed yearly | Searches | Bulk list size | API |
|---|---|---|---|---|---|
| Pro | 49 USD | 24 USD per month | 150 per month | none | none |
| Investor | 149 USD | 74 USD per month | 750 per month | up to 200 names | none |
| Broker | 399 USD | 199 USD per month | 3,000 per month | up to 1,000 names | 5,000 calls per month |
Each name in a bulk list counts as a search. Limits count usage in the billing period, and nothing is unlimited. The domain comps API page has the details.
How SoldComparables handles a name with no comps
This is the situation where model estimates and comp-based estimates differ most. A model will usually produce a figure for any input. We do not. If the closest recorded sales are weak matches, the confidence drops and the range widens. If there is nothing comparable above 1,000 USD, you see that plainly. For a hand-registered brandable, that answer can save you a renewal fee or an overpriced listing that never sells.
You can try it without an account: one search per day shows three comps and the estimate range. After signing up you get one one-time full report with every comp and the rationale. There is no free plan.
Check a name with the domain appraisal tool, or see pricing for bulk and API plans.