EBO.com
Menu

Editorial policy

Effective September 17, 2026. Last updated October 10, 2026.

This policy describes how EBO2.com chooses sources, reads clinic websites and published papers, checks what it records, writes and corrects its guides, and keeps its independence. It describes what the site does today. Where a step is done by a script, we say what the script checks and what it cannot check.

Independence and funding

EBO2.com is independent. It is not owned by, funded by, or affiliated with any clinic, device manufacturer, or ozone-therapy trade group. It runs no ads and no affiliate links, sells no listings, placements, rankings, or badges, and accepts no payment for coverage or for testimonials. The site is paid for by its operator, EBO2.com.

Clinics are listed free, and nothing a clinic does or pays changes where its listing appears, which badges it shows, or what our guides say about it. If we ever offered a paid feature for clinics, we would first change this policy and our terms, show the price and terms to a clinic before it agreed to pay, and label the feature wherever it appears. No payment would buy a "Verified" badge.

How we choose and grade sources

Every guide lists its sources at the end, and each source carries a label so that you can weigh it:

We prefer primary sources: the regulation rather than an article about it, the paper rather than a press release, and the regulator's own page rather than a news story. A clinic's website is a source only for what that clinic says about itself. When a guide cites a paper that is in our research library, the source list links to our notes on it.

We do not give studies a score. Each study in our library is labeled by its type, from randomized controlled trial to case report, laboratory study, or review, with the number of people studied when the paper states it and what limits the result. Studies are grouped by what they tested: EBOO as given to people, safety and adverse events, other ozone methods, and background science. Each group carries a note on what it can and cannot show. The type, the size, and the limits tell you more than a grade would.

How we describe evidence

We name the type, the size, and the main limitation of any study we describe, and we report what it measured and found, with attribution. We do not describe a result as more than the study shows: a change in a blood marker is reported as a change in a blood marker, not as a benefit to patients. Where no controlled evidence exists for a use, we say so before describing the rationale that clinics or researchers propose. A result for another ozone method is labeled as such, because it does not transfer automatically to EBO2.

Every guide states that ozone therapy is not FDA-approved to treat any condition, and our medical disclaimer appears at the end of every guide and study page. We do not use words such as "proven" or "cure" for any treatment, and an automated check enforces that in our guides, as described below.

How the research library is built

The research library holds 78 published studies: every paper on EBO2 (also called EBOO) that we have found, plus the ozone-therapy research that clinics most often cite and the published reports of harm. Every entry is a paper published in a journal and indexed in PubMed. We leave out conference abstracts, preprints, editorials without data, marketing white papers, and unpublished clinic data.

Each entry records the paper's authors, journal, year, and identifiers, its study type, and the number of people studied and the population when the paper states them. Two short notes are ours: what the authors report they found, and what limits the result. We write the limits from the authors' own stated limitations and from plain facts about the design, such as a single patient, no comparison group, one center, an animal study, or a reading of the abstract alone. The notes carry no praise or doubt. Each entry links to the paper on PubMed, and a study's page lists any guides that cite it.

Study cards

For each paper in the library we also keep a study card: what the paper reports, in structured form. A card holds the design, the setting, each study group with its size, the procedure's parameters exactly as the paper states them and in its units (ozone concentration, blood volume, session length, number of sessions, blood flow, anticoagulant, and filter), the outcomes it reports, including those that showed no difference, adverse events, follow-up, and the funding or conflict-of-interest statement. Every fact carries the paper's exact words, up to 40 words, and the page they came from. A field stays empty when the text we could read does not say, and we never fill a gap from memory, from our own notes, or from another paper.

How we read the papers

Papers are read from PubMed, PubMed Central, and Europe PMC, and occasionally from the publisher's own page. Newer cards are read through NCBI's E-utilities, the public programming interface for PubMed and PubMed Central, which returns the PubMed record and, for papers in PubMed Central's open-access collection, the full text. Of the 78 cards, 40 were read from the full text and 38 from the abstract alone, and each study's page says which, with the date it was read. A card read from the abstract cannot report what only the methods section says.

Errata and retractions

For every card so far we have also fetched the paper's PubMed record through E-utilities. It lists any erratum, retraction, or expression of concern that PubMed links to the paper, and a card records any such notice in its exact words. None of the 78 cards records one today. A comment on a paper is not a notice, and we do not record it as one.

How the quotes are checked

The cards hold 885 quotes. We save the text of every page read for a card, and a script checks that each quote appears in that saved text, allowing only for differences of typography, capitalization, and formatting, and that each study group's size appears in its quote. A card is not added until its quotes pass. Each card appears on its study's page, with links to the pages it was read from, and in the studies file on our data page.

How clinics are listed and verified

A clinic is listed when public sources or the clinic itself indicate that it offers EBO2 or EBOO. We find clinics through web searches and through requests that clinics send us. Other directories can help us find a clinic's name, but we never use them as a source: a listing is built from the clinic's own website, and the page that shows the service is recorded as its source. A site that mentions only "ozone therapy," without EBO2 or EBOO, is not listed as offering EBO2.

A listing holds the clinic's name, address, phone number, website, the services it lists, the practitioners it names, a price if it publishes one, and the dates we checked. "Verified" means that on the date shown we confirmed, on the clinic's own website, that it offers EBO2 or EBOO, and the listing links to that page. "Last checked" is the last date we reviewed the listing. 174 of the 177 listings are verified today; the others are marked "Not yet verified."

Who checks a listing

Listing checks are done by our research process, not by a named editor. AI research agents, working under the rules above, read the clinic's own website, find the page that names EBO2 or EBOO as a service, and record that page and the sentence on it that names the service. The sentence stays in our records, and the listing links to the page. A "Verified" badge means that this check found the service on the clinic's own website on the date shown; it does not mean that a person on our team reviewed the listing, and no clinician checks listings.

How listings are ordered

Within each state, the directory's list of clinics, and the lists on state and city pages, show verified clinics first and then sort alphabetically, and nothing else affects that order. The table on state and city pages that sets clinics' stated facts side by side lists first the clinics that publish a price under the published-price rule below, then sorts alphabetically, so that the prices sit together and are easier to compare; it is not a judgment about any clinic. The clinic finder sorts by straight-line distance, with clinics at the same distance in alphabetical order, and the list of nearby clinics on each clinic page shows other clinics in the same state, nearest first. Tables of every published price run from the lowest price to the highest. No ordering anywhere on the site takes payment, or any relationship with a clinic, into account.

Details from the clinic

A clinic can claim its listing at app.ebo2.com with a confirmed email address on its own website domain, and then set its phone number, public email address, booking link, and opening hours, which appear within a few minutes, and propose its services and a short description, which a person on our team checks first against the same rules as our own text: no claim that a treatment cures, treats, or prevents disease. On the clinic's page these details are marked as coming from the clinic, and they never change what we report from the clinic's own website. The services a clinic lists for itself show on its page; our guides, reports, and downloads keep counting the services we recorded from its website. Claiming is free and does not change a listing's position.

How listings are rechecked

Our listing procedure calls for a recheck of every listing that reaches 90 days since its last check. We reopen the page recorded as the verification source. If the clinic still lists EBO2 or EBOO, we update the verified and last-checked dates; if it does not, we remove the listing or mark it not verified. A script lists every listing past 90 days, and our automated tests fail once any listing passes 150 days.

The site's build enforces a hard limit. If any listing's last check is more than 180 days old, the build stops with an error, and no new version of the site can be published until that listing is rechecked or removed. The check blocks publication; it does not take a listing down from the version already online. The oldest check in the directory today is from September 16, 2026, so the next rechecks fall due by December 15, 2026.

What clinics say: the research pass

Separately from the listing check, a research pass reads each listed clinic's EBO2 page and records what the clinic states about the service. The current records were read between September 28, 2026 and October 4, 2026. Where the EBO2 page does not state a price or how a session works, the pass may read a few more pages on the same site, such as a pricing page or a page of questions and answers, up to six pages for one clinic in all. It never uses review sites, social media, other directories, or any site other than the clinic's own.

The pass records the single-session price, packages, session length, the device the clinic names, add-ons, who supervises sessions, what happens before a first session, the course the clinic recommends, and the clinicians it names. Each fact carries the clinic's exact words, from one sentence or phrase of at most 40 words, and the page they came from. A fact the pages do not state stays empty: the pass never infers, averages, or fills a gap from another clinic.

Who reads the pages, and how the quotes are checked

Clinic websites are read through a third-party web-reading service, and the reading and copying are done by software, including AI research tools, working under the rules above. When the service cannot load a clinic's page, an AI research tool reads the page in a web browser instead; where the text is copied out of the browser rather than saved by it directly, every line of the copy is checked against the live page before any fact is taken from it. That is one reason every fact is tied to the clinic's own words.

To check that what we show is still on a clinic's pages, a script re-reads each clinic's pages in a cloud browser service and lists every recorded quote that is no longer there, every page that no longer mentions EBO2 or EBOO, and every page that does not load. A fact the check flags is read again and checked as above before its record changes.

We save the text of every page read. A script checks that each quote appears in the saved text of that clinic's pages, allowing only for differences of typography, capitalization, and formatting, and that each price, package size and total, add-on price, course length, and session length appears in its quote. Where our own tools saved a page (papers, and clinic pages read in a browser), the script also confirms that the saved text has not been changed since. A record is not added until its quotes pass. The 177 records hold 1,237 quotes. The saved copies are kept for these checks and are not published.

What a clinic page shows

A clinic page shows the facts the research pass recorded, each with a link to its source and, where the wording adds something, the clinic's own words. "Read on" gives the date the pages were read. For a clinic whose pages we read, four key facts are marked "Not stated on the pages we read" when those pages do not state them: the single-session price, the session length, who supervises, and the device. That phrase means only that the pages we read do not say; it does not mean that the clinic does not do it. Where we could not read a clinic's pages, its page says so, and we never describe that clinic as not stating something. 3 of the 177 records are of this kind.

Facts are shown in plain terms. When a clinic's wording for a value carries a health, safety, or regulatory claim, the page shows "Described on the clinic's page" with a link to the clinic's page instead of repeating the wording, and the clinic's own quotes appear only when they carry no such wording. A clinic page is not offered to search engines unless the research pass has read the clinic's pages.

What we never show for a single clinic

The research pass also records three things that we publish only as counts across all clinics, never for a named clinic: how the clinic's pages word the FDA status of EBO2 or its equipment, whether its EBO2 pages state that the procedure treats or helps a named condition, and whether they show patient testimonials. None of these appears on a clinic's page, in the clinic finder, or in our downloads.

The clinic transparency report gives the counts for FDA wording and for health benefits, sorted from each page's own words into kinds such as detox, infection, autoimmune conditions, and cancer. A general statement such as "supports wellness" is not counted as a claim. We publish these only as counts because each is a judgment about a named business's marketing. Counting them shows what is common, and what to look for on a clinic's own pages, without characterizing any one clinic. A statement on a clinic's page is a claim, not evidence; for what the studies show, read our guide on whether EBO2 works.

Regulatory records

The regulatory tracker holds 33 records of what regulators, courts, legislatures, and professional bodies have said about ozone therapy, EBO2, ozone devices, the filters used in these circuits, and health claims. Each record comes from a primary source: the issuing body's own page or document, such as an FDA warning letter, a section of the eCFR, an FTC letter, a state board's statement, a court decision, or a regulator abroad. News stories, clinic blogs, and commentary can lead us to a document, but they are never the source.

A record holds the document's own date, the body that issued it, the kind of document, a neutral summary in our words, and a key quote copied exactly from the document, up to 40 words. Documents in another language are quoted in the original and summarized in English. A record names a company or a person only when the document itself does and the document is public, as a warning letter is. Its status, current, superseded, or historical, is as of the date we checked the source. A script checks that each key quote appears in the saved text of the regulatory sources we read. We record what a document says, not what we think of it: a warning letter states an agency's position on the date it was sent, and it is not a court finding.

What our checks do not cover

Our checks confirm that the words we quote are on the pages we read. They do not confirm that those words are true.

Figures, reports, and downloads

Every figure in our reports, tools, guides, and answers to common questions is computed from these records when the site is built, so a figure changes when the data does. Each says how many clinics or studies it rests on and when the data was read. A share always names what it is a share of, and a clinic whose pages we could not read is left out of a share rather than counted as saying nothing. Our reports give a median or another summary only when at least 5 clinics publish a figure, and below that they list the values. The price check places a quote only when at least 5 clinics publish a single-session price.

The published-price rule

Every count, share, median, and range of published prices on the site is built from one set of prices, wherever it appears: the home page, the reports, the figures in our guides, the tools, the state, city, and clinic pages, and llms.txt. The set starts from the clinics whose EBO2 page our research pass could read. Among them, a clinic counts as publishing a price when that page states a price for a single session, with the exact words saved and checked by the quote script. Two kinds of price stay out of both the count and the total it is a share of. A price recorded only when the listing was checked is left out because its wording was never checked against a saved page. A price of a clinic whose EBO2 page could not be read, or has not been researched yet, is left out because counting it would bring that clinic into the total only because it has a price, which would overstate how many clinics publish one.

A price recorded when the listing was checked still appears where it belongs, labeled by its source: on the clinic's own page, which may compare it with the median; in the clinic data files, whose price_source field says which kind it is; and in the clinic finder, which marks it "From" and includes it when a reader filters for clinics with a price. The transparency report's "publish a price" row uses the same test as every price figure, so the counts agree.

The downloads

The files on our data page are rebuilt from the same records every time the site is, and each states when its data was read. They are published under the Creative Commons Attribution 4.0 International license (CC BY 4.0), which lets you reuse them as long as you credit EBO2.com, link to the data and to the license, and say whether you changed them. The clinic files leave out how each clinic words FDA status, whether its pages state health benefits or show testimonials, clinicians' names, phone numbers, street addresses, and our research notes. They do give each listing's city, state, ZIP code, and map coordinates. The regulatory file, regulatory.json, names the party each document names, which can be a company or a person, and the study file names each paper's authors. The license is the site owner's choice and may change for future versions of the files; a change does not take the license back from copies already published under it. Our terms of use explain what the license covers and what it does not.

How the tools work, and what they never do

Our four tools are built from the same records as the reports, as of the date each tool's page shows. The records are built into each page when the site is built, and the tool does its arithmetic in the reader's browser. Each tool's page says how it works, which data it uses, how many records that is, when they were read, and what it cannot tell you.

What the tools never do

No tool scores, rates, or ranks a clinic, and none advises. The price check describes where a number falls and never calls a price good or bad. The planner adds up costs and never suggests a number of sessions or says whether a package is a good idea. The clinic finder sorts by distance alone and never recommends a clinic. The questions for your doctor asks; it never scores a reader's answers, ranks situations, warns anyone off, or says whether a reader can or should have EBO2. No tool takes payment or any relationship with a clinic into account, and none sends us what a reader enters, apart from the first digit of a ZIP code, which shows in the address of the one ZIP code table the clinic finder downloads, as our privacy policy explains.

How guides are drafted, checked, and updated

Our 29 guides are drafted with AI writing tools from the sources they cite. A human editor has not yet checked every claim in every guide against its cited source; each guide's closing paragraph says where its own review stands. Guides cite their sources by number, and every number points to an entry in the source list at the end of the guide.

Automated checks on every guide

An automated check reads the text of every published guide. Our automated tests run it, and it fails when a guide:

A source that is listed but never cited produces a warning rather than a failure. The check reads a guide's main text, not its takeaways or its questions and answers, and it cannot judge whether a sentence is true: that takes a person reading the guide against its sources, and, as above, a human editor has not yet done that for every claim. It skips drafts, which are not published.

Numbers from our clinic data

The numbers our guides and answers state about clinics (how many publish a price, what they claim, what their pages say about who supervises a session) are not typed by hand. Each is computed from our clinic records when the site is built, with a written method, so the sentence and the "From our data" figures below it always agree. Where a sentence's wording rests on a number, such as "most clinics", "fewer than half", or "one clinic", a check stops the site from being published when the data no longer supports that wording, and the sentence is rewritten first. Some counts come from reading what each clinic's saved page says, such as which prices come with a condition; our research process (AI research agents reading the saved pages) classifies each clinic, and when a clinic's record changes, the site cannot be published until that clinic has been classified again.

Updates and illustrations

Each guide shows when it was first published and when it was last updated. When we revise a guide in substance, its Updated date changes and its first-published date stays the same. The drawings at the top of guides are made with AI image tools from written prompts. They show objects, never people, and they illustrate a topic rather than document anything.

Medical review

None of our 29 guides has been medically reviewed yet. A "Medically reviewed by" line appears on a guide only after a named, licensed clinician has reviewed and approved that text, and it shows the reviewer's name, credentials, and the date of the review. The site cannot show the line without both a reviewer and a review date, and a reviewer's profile must list credentials. Review is not an endorsement of EBO2, of any clinic, or of ozone therapy. Our medical review page explains what a reviewer checks, what review does not mean, and how reviewers are chosen and paid.

Conflicts of interest

A conflict of interest is any tie that could pull our coverage toward a clinic, a device, or a conclusion. The site has none of the usual ones: no advertisers, no affiliate partners, no paid placements, and no funding from clinics, device makers, or ozone-therapy groups. Medical reviewers will be paid a flat rate for their time that does not depend on what a guide concludes.

If anyone who writes, checks, or reviews a page has a financial or professional tie to a clinic, a device maker, or an ozone-therapy organization that the page discusses, we will say so on that page. A clinic that asks us to correct its listing is treated the same as a reader who reports an error, and no clinic can pay to have a guide changed or a listing moved.

Corrections

Report an error through the contact form and choose "Report an error or request a correction." Each clinic listing also has a "Tell us" link to the same form. Tell us the page, what is wrong, and, if you can, a source that shows the right information.

If we cannot confirm a change, we tell you and leave the page as it is until we can. If we disagree, we explain why.

How a clinic can dispute a listing

A clinic can claim its listing, correct it, add a short description shown as "Provided by the clinic," or ask us to remove it, through the For clinics page. Before we act on a request as coming from the clinic, we confirm that the email address it came from is on the clinic's own website domain, or we ask for a reply from an address on that domain. We apply a change to an address, phone number, email address, practitioners, or description only when the same information appears on the clinic's website, because that is the source every reader can check. If it is not there yet, we ask the clinic to publish it first.

A clinic can also tell us that a guide, a fact on its listing, or a quote misstates what its pages say, for example because it has changed a page since we read it. We review every such message, correct what is wrong, and explain why if we disagree. A listing is removed when the clinic asks, closes, or stops offering EBO2, and its page then stops working. Disputing a listing costs nothing, and it does not change how the listing is ordered.

What we never publish

We also decline listing text from a clinic that claims a treatment cures, treats, or prevents disease, includes testimonials or reviews, says or implies that ozone therapy is FDA-approved, or misrepresents anyone's credentials. When we quote a clinic, its words appear as a quote, attributed to the clinic, with a link to its page, and our screen keeps health claims out of the facts shown on its listing.

Our research tools and other sites

When our own scripts fetch a web page or a paper directly, they identify themselves as "ebo2.com-research," with a link to this page. They read only the pages of well-known public institutions: government agencies, intergovernmental bodies, scholarly indexes such as PubMed and Europe PMC, journals and publishers, and professional and standards bodies. Papers are read through NCBI's E-utilities, NCBI's public programming interface. The scripts read one page at a time with a pause between requests, and they do not try to get past bot checks, logins, or paywalls: a page that refuses them is recorded as unread. They do not fetch clinic websites directly: those are read through a third-party web-reading service or in a web browser, as described above, and the script that re-checks clinics' pages opens them in a cloud browser service. That browser loads a page as an ordinary browser does, one page at a time; it never solves a CAPTCHA or gets past a bot check, and a page that refuses it is recorded as unread. If you run a site our tools have visited and have a question, or would like them to stop, use the contact form.

Contact

Questions about this policy: use the contact form or write to EBO2.com, 1968 S. Coast Hwy, #1921, Laguna Beach, CA 92651, United States.

Change history