src/content/company.ts for a fact, the site environment for an address - and this banner goes when the last of them is answered.Five honest alternatives, and what each one costs you.
Nobody buys provider data because they read a comparison page. You are choosing between a free registry download, a cheap scraped list, a contact database you may already pay for, a claims data licence, and building it yourself - and each of those is the right answer for somebody. This page says what each one is good at, what you take on with it, and where we are the wrong choice. No competitor is named here, and no claim is made about anybody else's product that we could not stand behind.
Six things, and every one of them is checkable.
A date on every field, not on the file
Identity and specialty are refreshed on one schedule, affiliations and licences on another, a work email continuously - and each field says when it was last confirmed. You quote that date, decide what is fresh enough for what you are doing, and stop treating a record as uniformly current.
A removal that keeps working
When somebody asks to be left alone it applies to search, pages, counts, exports, append and the API, for every customer, and it survives the weekly rebuild because the record is marked rather than deleted. It is also the reason we do not sell files: a removal cannot reach a copy somebody already owns.
Contact details metered and audited
A work email is opened one record at a time against an allowance and written to an audit log. Home address and cell phone are contract-only, matched to the named provider, and never in an export or a list. You can see what your own team did, and so can we.
The organisations, and who runs them
Certified facilities with beds, ownership, chain and published ratings, the systems above them, and the named people who run them by role - counted in public, named for signed-in customers. This is the part a registry download does not contain and a generic database was never built for.
An API and an append tool, not a file drop
Ask the question you have today rather than the one you anticipated when a file was cut. Append your own list and get the identifier, specialty, location and standing back against your own rows. Exports exist, with an allowance, for work that genuinely needs a file.
What they do, not only who they are
The procedures a provider billed, code by code, and the drugs they prescribed, by brand and by generic — from published Medicare data for 2024. Searchable by name rather than by code, so “total knee replacement” and “Eliquis” are queries. It is Medicare fee-for-service only and covers about one in seven clinicians nationally, and far more of the specialties that bill it most, which is less than a claims licence and a great deal more than a specialty label. The limits are published, per specialty, rather than discovered.
Coverage you can check before you buy
Counts per specialty and per state are published on the list pages - records, with a work email, with a trusted-tier email, with a practice phone. Check your own segment rather than taking a national total on trust.
Read live from the same API a customer calls, as this page was served: 9,469,440 provider records, 48,350 certified facilities, 24 systems and chains with a page of their own, 143,501 leadership contacts counted by role.
What each one is good at, and what you take on with it.
A public registry download
What it has going for it. It is free, it covers the whole country, and it is the same public record we start from. If identity, specialty and a practice address for a known list of providers is genuinely all you need, download it and keep your money.
What it costs you.
- The file has one date. A record does not tell you which of its fields was confirmed when, so every value is as old as the slowest part of the file.
- There is no work email in it, and no way to tell a live address from one that was filed years ago and never revised.
- Certified organisations, their beds, ownership, chain and published ratings are a different problem entirely, and the people who run them are not in any filing.
- Somebody has to load it every week, reconcile the changes against what you already had, and keep doing that for as long as you use it. That work does not stop being work because the file was free.
- Nobody has asked to be left out of it on your behalf. When a provider asks us to stop, that holds for every customer, and it keeps holding through the next refresh.
A scraped list
What it has going for it. It is cheap and it arrives immediately, and if you only need to test whether a market exists at all it will tell you that.
What it costs you.
- Ask where each field came from and when it was confirmed. A scrape has no answer, because the answer was never collected - and the page it was taken from is not a filing anybody stands behind.
- Ask what happens when somebody asks to be removed. There is no mechanism, so the answer is nothing, and the same details come back in the next copy of the list.
- Ask who is liable for how it was collected and what you may do with it. On this platform that is in the terms, on a page, before you buy.
- A list you cannot audit is a list you cannot defend, and it is your outreach that arrives at the wrong person, not the seller's.
A generic B2B contact database
What it has going for it. For reaching software buyers across every industry, a broad contact database is the right tool, and one of them is probably already in your stack. Some of them are very good at company hierarchies and at email.
What it costs you.
- Healthcare specialty is not a job title. A clinician's specialty here is the taxonomy they filed, with the classification it sits under, so asking for a classification matches every specialty beneath it - rather than matching whatever somebody typed into a profile.
- The ten-digit identifier is the join key. It is what lets you append to your own file, deduplicate against it, and go back to the same provider next quarter and know it is the same person.
- Ask whether the record tells you if a provider can be billed for at all - exclusions, opt-out and revocation flags. It is the cheapest disqualification there is, and it is on the record here before you spend anything.
- Ask what it holds about the organisation: whether it is certified, how many beds, who owns it, which chain it reports under, its published ratings, and who runs it by role.
- Ask how a specialty count for your own segment is published before you buy. Coverage per specialty and per state is on the list pages here, so you can check your segment rather than take a total on trust.
A claims data vendor
What it has going for it. On breadth of clinical activity they beat us and we are not going to pretend otherwise. All-payor claims see commercial, Medicaid and Medicare Advantage volume, not just Medicare fee-for-service; they see it for very nearly every practising clinician rather than the roughly one in seven whose Medicare volume is published; they carry years of history, so they can tell you a procedure count is rising; and they see it sooner. If your question is "who is doing more of this than last year, across all payors", buy claims.
What it costs you.
- We do hold what a provider bills and prescribes, and it is a real answer to "what do they do": procedures by code and drugs by brand and generic, for 2024, filterable by name and readable per provider. 1,144,245 of 7,492,373 active individual providers (15%) have a published procedure line and 1,135,392 have a published prescribing row, measured 2026-10-05. That is the national share, and it is much higher in the specialties that bill Medicare Part B most - the data page publishes it per specialty, so you can check your own segment. Where that overlaps your target list, it does the job.
- Ask what it costs and what you may do with it. Licensed claims is a different order of price, and the licence usually restricts marketing use of the derived list - which is the use most people were buying it for. Ours is published data, and what you may do with it is on a page before you buy.
- Ask what you get for the clinicians the claims file is thin on, and what the contact data looks like. A claims record tells you what somebody did; it does not reliably tell you their current practice address, a confirmed work email, the organisation they bill under or who runs it. That is the part of the job you would still be solving.
- Ask whether a removal request reaches it. When a provider asks us to stop, it holds across search, pages, counts, exports, append and the API, for every customer, and it survives the next refresh.
- The honest division: claims is the better answer to “who is doing more of this”, and we are the better answer to “who does this, who exactly are they, and how do I reach them”. Plenty of teams need both, and this is the cheaper half.
Building it in-house
What it has going for it. It is a real option, and if you have a data team and need only identity and specialty, it is not a hard build. You would own it, and nobody would meter you.
What it costs you.
- Identity is the easy part. Reconciling one provider across filings that disagree, and keeping that stable as they move practice, is where the time goes.
- Work email cannot be read off a filing. It has to be confirmed, scored, and re-confirmed, and an unscored address has to fail rather than default to usable - which is the decision most in-house builds get wrong once and then live with.
- Removals have to survive a rebuild. If your pipeline recreates the table each week, a deleted row comes back, and you will not notice until somebody writes to you a second time.
- Then the parts nobody scopes: certified facility data, ownership and chains, leadership by role, an audit trail of who revealed what, and the ability to answer a customer asking where a field came from.
- Keeping it current is the whole cost. The first load is a fortnight; the fifty-second load is the job.
Four cases where the answer is not us.
A comparison page where the answer is always the seller is an advertisement. These are the requirements we cannot meet, and knowing them now is cheaper for both of us than finding out in week three.
You need anything about patients
There is none here, in any form, and there never will be. Not identified, not de-identified, not aggregated. If your question is about treatment, outcomes or claims at the patient level, this is the wrong platform and no plan changes that.
You need all-payor claims, or any history
What we hold is published Medicare fee-for-service data for 2024 - one year, no earlier years - covering about one in seven clinicians nationally, because low-volume providers are removed from the published file before we receive it. If you need commercial, Medicaid or Medicare Advantage volume, near-complete clinician coverage, or a trend across years, a claims licensee has it and we do not. Saying so costs us a deal occasionally and saves us the ones that would have gone wrong in month two.
You want consumer marketing data
Provider records are not mixed with a consumer file, blended into a count or shipped inside a provider export. Provider-matched home address and cell phone exist under a signed contract with a stated permissible use, revealed one provider at a time, and are never in a list.
You want to buy a file and resell it
No plan carries redistribution or resale rights, exports are traceable to the export that produced them, and contact details are metered rather than handed over in bulk. If your model needs a file you own outright, we are not it.
You need a handful of providers you can already name
Search is open without an account, and a record page carries identity, specialty and practice location. If that is the whole job, do it for free and do not talk to us.
[OWNER: whether any customer may be named on a public page, and in whose words - no name goes on this site without their written agreement]
What people ask when they are comparing.
Why not just use the free public file?
If identity, specialty and a practice address for a list you already have is all you need, use it. What you are paying for here is the part the file does not contain: a confirmed date on each field rather than one date on the whole file, a work email that has been confirmed and scored, certified organisations with their ownership and published ratings, the named people who run them, a removal that holds through every refresh, and an API and an append tool instead of a weekly load you maintain yourself.
Their data says what a doctor does. Yours only says who they are, doesn't it?
Not any more, and not entirely before. We hold what a provider billed Medicare for, procedure by procedure, and what they prescribed under Part D, drug by drug, for 2024 - and you can filter on it by name rather than by code, then read the detail for a provider. Two things are true about it that we would rather you heard from us: it is Medicare fee-for-service only, so commercial, Medicaid and Medicare Advantage volume is not in it, and it covers about one in seven clinicians nationally - and a good deal more of some specialties than others, which the data page publishes per specialty, because low-volume providers are removed from the published file before we receive it. An all-payor claims licence is genuinely broader and genuinely sees history we do not have yet. What it usually does not give you is the current practice address, the confirmed work email, the organisation and the removal discipline you need to act on any of it. The two are complements, and this is the cheaper half.
How do I check your data is better than what I have now?
Take twenty providers you know the truth about and look them up. Search is open without an account and needs no conversation with us. Compare name, specialty, practice address and phone against what you know, and check the confirmed date on each field. Twenty records you can verify will tell you more than any figure on any vendor's page.
Why is there no file of contact details to buy?
Because a database that can be emptied in an afternoon cannot be kept honest. Contact details are opened one record at a time, against an allowance, and written to an audit log, so we can see how the data is being used and act when it is being abused. It also means somebody who asks to be removed is actually removed, rather than removed from our copy while a hundred files of it are already elsewhere.
Is an API really better than a file?
For most of what people buy provider data for, yes: you ask the question you have today rather than the one you anticipated when the file was cut, your own records are appended in place, and you are reading the current state rather than a snapshot ageing in a bucket. Exports exist for the work that genuinely needs a file, with an allowance behind them.
Check it on twenty providers you already know.
Search is open without an account and without a conversation. Look up people whose details you can verify, check the confirmed date on each field, and decide from that rather than from this page.