Skip to content

Legal

Intellectual Property, Copyright and Trademarks

Who owns what, how we handle third-party material, and how to complain.

Last updated 28 July 2026

1. Our material

1.1 What is ours. This website, its design, code and copy; the rubric and the criteria behind each dimension; the section structure of each report; the citation grammar findings are written in; the report templates; the published sample reports; and the names in section 8.

1.2 What we do not claim, and what you may rely on. A report's text is produced by automated systems, and no person reviews, edits or approves it before it reaches you. The law of some jurisdictions grants no copyright in material generated that way, and prompting a system is not by itself authorship. So we do not claim that the text of your report is protected by copyright, and we do not claim that any rights we hold in it are exclusive. That is a statement about what subsists in the text. It is not a warning about what you are buying, and it is not an allocation of risk to you. What we do say, and what you may rely on: we are entitled to supply your report to you, no other person's permission is needed for us to do so, and we know of no third party's right that prevents you using your report for the purposes clause 8 of the Terms of Service permits. Where a source is quoted inside your report, clause 3.4 says where the rights in that passage sit and clause 1.6 says what happens if one of them gets in your way.

1.3 The protection around a report is contractual, it does not depend on ownership, and the licence is not unconditional. The restrictions on what you may do with a report – no publication, no resale, no syndication, no use to build a competing product or dataset – are in clauses 8 and 9 of the Terms of Service. They bind you because you agreed to them, not because of any right of ownership, and they would still bind you if a court found that no copyright subsists in the text of a report.

Each of those restrictions is a separate and independent undertaking. If one of them is held not to bind you, the others are unaffected and continue to apply. Two of the protections around a report do not depend on our contract with you at all: the material in 1.4 is confidential whether or not you have agreed to anything, and rights in third-party material quoted inside a report belong to their owners under 3.4 and were never ours to license to you.

Facing the other way, you keep your report and may use it for the permitted purposes for as long as the licence in clause 8 of the Terms of Service runs, and while it runs we will not assert a copyright claim against you for using it that way. That licence, and with it your right to use the report or any copy of it, ends in these cases and only these: you are refunded in full, whether you asked for the refund or we made it without being asked; you withdraw from the purchase under the Refund Policy; we withdraw the report because its subject objected, or withdrew the agreement described in 2.4, or because the law or a court requires it; or the licence is revoked for breach under clause 8 of the Terms of Service, which includes telling us under 2.4 that you had a person's permission when you did not. A payment reversed by your bank or card issuer has the same effect as a full refund, from the moment the reversal is final. In each of those cases you must stop using the report and delete the copies you control, as the Refund Policy and clause 10 of the Terms of Service require. Using it after that is a breach of contract whether or not any copyright subsists in the text, and that is exactly why 1.2 costs us nothing.

1.4 What we treat as confidential. Our prompt library, our model roster, our routing configuration and the parts of the pipeline we have not published are trade secrets, kept unpublished and under access control.

1.5 What is published stays published. The six dimensions, the criteria text, the five-point judgement scale with its fixed points table, and the plain unweighted mean that produces the headline number are all published, at /intelligence-score and in AI Transparency and Limitations. Publishing them neither discloses nor waives protection in anything in 1.4. We would rather publish the whole scoring rule and keep the plumbing than publish half a rule and call it transparency.

1.6 Consumer rights this notice does not affect, and what we do if a third party's right gets in your way. If you are a consumer, nothing here limits your non-excludable rights, and in particular your right that digital content supplied to you be free of any third party's right that would prevent you using it as agreed. Clause 1.2 states what we can and cannot claim to own. Clause 3.4 states where rights in quoted material sit. Neither is a term shifting that risk onto you, and neither may be read as one.

If a third party's right prevents you using your report in a way clause 8 of the Terms of Service permits, write to [email protected]. We will remove or replace the material that right attaches to and regenerate your report, at our cost, and we will not ask you to establish the right first. If that does not restore the use clause 8 permits within a reasonable time, or cannot be done without significant inconvenience to you, you are entitled to a reduction in the price or to end the contract and be refunded, and the choice between those two is yours and not ours.

Those remedies come from the law rather than from us. Nothing in this notice replaces, conditions, delays or exhausts them: you do not have to complain under section 10 first, and nothing here makes your remedy depend on our agreeing with you. The Refund Policy says how the money is returned.

2. Your material

2.1 We claim nothing you submit. A brief is a subject name, a link and an optional line of context. It stays yours. The licence you give us to use it is in clause 5 of the Terms of Service and is limited to producing and delivering your report and keeping the records described in the Data Retention notice.

2.2 Your context is never quoted. What you tell us steers where we look. It will never be treated as evidence, quoted in a finding, cited as a source, shown on a shared view or served in the report data. Clause 6 of the Acceptable Use Policy is the full statement of that rule.

2.3 You warrant that you may supply it. If what you submit contains someone else's copyright material, mark or confidential information, that is your representation to make and yours to be right about.

2.4 Naming a person is not a permission from that person, and we do not treat it as one. Where a report is about somebody other than you, the name, link or context you give us is information about a third party. Supplying it gives neither you nor us any right in that person's material, likeness or name.

A report about a natural person other than you will be produced only where that person has confirmed to us, in reply to a request we send them, that they agree to it. The request will describe what the report is and who asked for it. Your statement that you have that person's permission is evidence and carries the warranty and the indemnity in the Terms of Service; it is not a substitute for their confirmation to us, because a permission given to you is not a permission we can demonstrate, and a statement you knew to be untrue when you made it is a material breach that ends your licence under 1.3. Where no confirmation arrives within the window that request states, the order is cancelled and refunded in full, automatically, and no report is produced.

An agreement given that way can be withdrawn afterwards. How it is withdrawn, and what we do about a report already produced, are stated for the person who gave it at If a Report Names You; the effect on your licence is in 1.3. Who a report may be about is in the Acceptable Use Policy; the position of the person named is set out for them, in their own words, at If a Report Names You, and nothing in your contract with us limits it.(/legal/privacy) rather than this notice that states it. Second, confirm that a confirmation captured this way meets the demonstrability standard for the controller, and how long the record of it may be kept, which is a question for the Data Retention notice. Third, confirm that ending a buyer's licence for a statement under 2.4 that the buyer knew to be untrue is enforceable against a consumer, and note that this notice deliberately says nothing about whether the price is returned in that case: the Refund Policy is the only place that is answered, and it must be conformed to carry the position rather than leave it to be inferred from here.]

3. Third-party material in a report

3.1 What a report contains that is not ours. Source links, source titles, publication dates, short quotations from a source, and our own description of what a source says. Our description is our statement and not the publisher's; the difference between an attributed fact and our own assessment is set out in clause 2 of Report Accuracy and Public Data.

3.2 Three different acts, three different questions. We separate them deliberately, because they are not answered by the same rule and folding them into one sentence about mining would hide the two that are hardest.

ActWhat it isWhere it is answered
Reading and analysing a public pageAutomated analysis of content we did not writeSection 6
Keeping what we read as evidenceRetention so a buyer can check a finding against its sourceClause 6.6 and the Data Retention notice
Reproducing words from it in a paid reportQuotationSection 4

3.3 What is never reproduced. A full article or a substantial part of one. Images, logos, charts or any graphic from a third-party site. Paywalled or credentialed content in any quantity. Song lyrics or poetry in any quantity, because a short quotation cap does not save a taking that is substantial relative to a short work. Source code. Verbatim search result blocks. Another vendor's report, in whole or in part.

3.4 Rights in quoted material stay with their owner. The licence in clause 8 of the Terms of Service covers your report as a whole, including the quoted passages inside it, for the purposes that clause permits: reading it, printing it, storing it, and showing it to the people it allows. What the licence does not give you is any wider right in the quoted material itself. You may not lift a quoted passage out of your report and publish, resell or syndicate it, and if you use one for a purpose the licence does not permit, that is your act, governed by the source's own terms and not by ours. If a third party's right prevents you using your report for a permitted purpose, tell us: clause 1.6 applies and we do not put that risk on you.

4. Quotation limits, and the gate that applies them

4.1 The caps we will apply. Where a report quotes a source, these will be the limits, from the first delivered report:

LimitValue
Words in any one quotation25
Quotations from any one source document2
Total quoted words from any one domain, in any one report200
AttributionEvery quotation, with a link that resolves to the source

Clause 4.2 says what these are and what they are not. Clause 4.3 says what enforces them and what does not enforce them yet.

4.2 What the caps are, and what they are not. They are a limit we place on our own conduct when we collect material and compose a report. They are not a specification of the document you buy. A report is not defective, deficient or non-conforming because a quotation in it is shorter than a cap allows, because a source was dropped to stay inside one, or because a cap was reached at all. What you are entitled to receive is described in clause 2 of the Terms of Service and in clause 1.1 of Report Accuracy and Public Data; this section adds nothing to that description and takes nothing away from it.

Where a quotation would breach a cap, or its source link does not resolve, it will be removed in full rather than trimmed, because a quotation trimmed by a machine is a quotation nobody decided about. The finding it supported will either be re-derived from the same source without the quotation or dropped, and where dropping it means a source is not used, the source is named as unavailable under 6.3. A report fails, and is refunded automatically under the Refund Policy, only where what remains cannot meet the requirements in clause 1.1 of Report Accuracy and Public Data.

If you think a quotation in your report has taken too much from a source, tell us under section 10 and we will shorten or remove it and regenerate the document at no cost to you. If you own the source, section 10 is your route and 4.6 states what our answer to you is not.

4.3 The state of this section, as at the effective date above. The caps in 4.1 will be enforced by an automated check running ahead of delivery, which counts the words in each quotation, the quotations taken from each source document and the total quoted words per domain, and requires every quotation to carry a source link that resolves. A report that exceeds a cap will fail that check and will not be delivered.

That check is not built. Until it is, the caps are a rule of composition with no automated gate behind them, and we say so here rather than let you assume otherwise. Building it is a condition of the first sale. This clause is reviewed at every release of this notice and is removed in the release that makes it unnecessary, with the removal dated in the change record.

4.4 Counting, so a publisher gets a number. Each report will record how many words were quoted from each domain. A publisher who asks how much of theirs we have used will get a figure rather than an argument. That record is produced by the check in 4.3 and does not exist before it.

4.5 A do-not-quote list. A domain we are asked to stop quoting goes onto a list we keep from the effective date of this notice. From the release in which the collection stage exists, a domain on that list will be excluded when material is collected rather than filtered out of a finished report, so a publisher who tells us once will not have to tell us twice. Clause 6.4 is the route, and a domain added now is on the list before the first report runs.

4.6 What the caps are not. They are a limit on how we build and operate the product. They are not a legal opinion, they do not decide whether any particular taking is permitted, and pointing at them is not our answer to a rightsholder who says we went too far. If you own a source and think we did, section 10 is the route and we will look at it on its own facts.

5. What we never do

5.1 How we reach a page. We read only what a browser can reach without credentials, and we never buy files about people. Clauses 3.1 and 3.2 of Report Accuracy and Public Data are the full statement of that rule and this notice does not restate it differently.

5.2 We never build a copy of somebody else's site. We do not republish, mirror or syndicate a source's content, and we do not make what we have fetched available to anyone as a corpus, dataset, feed or searchable archive. We do not sell evidence and we do not license it on.

5.3 We never take the visual identity. No logo, no wordmark typeface, no brand colour scheme, no product imagery.

5.4 We train no models of our own on anything we collect or anything you submit. What we send to the search and model providers named in Sub-processors, and the conditions we will place on them, are in clause 11.4 of AI Transparency and Limitations, which is the authoritative statement, is narrower than a blanket promise about what those providers do, and governs if this clause is read more widely than it.

5.5 We never imply a relationship. We do not use anyone's mark in a way that suggests one, and we do not describe a subject as verified, certified, endorsed or approved by us.

5.6 Two limits that live in other documents and belong here too. Material revealing a special category of personal data is outside what we collect and outside what a report may contain, as clause 4.6 of If a Report Names You states. And a report about a natural person other than the buyer will be produced only after that person has confirmed to us that they agree to it, as 2.4 describes, and that agreement can be withdrawn afterwards. A rightsholder or a regulator reading this notice alone should not have to find either limit somewhere else.

6. Text and data mining, and reservations

6.1 What we rely on, and what it does not answer. Reading a public page and analysing it is not reproduction of what it expresses. For copyright and for database rights we rely on the exception in EU law permitting text and data mining for commercial purposes where the rightholder has not reserved the right in a machine-readable way.

That exception answers a copyright question and nothing else. It is not a lawful basis for processing personal data, it gives us no permission in relation to a person named on a page we read, and we do not treat a publisher's failure to reserve as something that person agreed to. Our basis for processing personal data, the balance we struck and how a person objects are in the Privacy Policy and in If a Report Names You. An objection made there stops a report being produced about that person whatever this section permits us to read, and clause 2.4 states the separate rule that a report about a natural person other than the buyer needs that person's confirmation to us before it is produced at all.

6.2 The reservations we read, and the ones we do not yet read. We recognise three mechanisms by which a rightholder may reserve: the robots instructions a publisher sets for our fetcher, a machine-readable reservation sent as a response header, and a reservation published in an ai.txt file at the site root. Our fetcher will read the robots instructions. A reservation sent as a response header, or published in an ai.txt file, will not be detected by our fetcher until the check for it is built, and we will not let you assume otherwise.

Three things follow, and we state all three rather than only the flattering one. First, a reservation made by any of the three mechanisms is effective against us from the moment we are aware of it, however we became aware of it, including by a message under 6.4, and we do not require it to have been made in a form our fetcher can read. Second, where we learn that we fetched a source that had reserved by a mechanism we were not reading, we delete what we took from it, remove every quotation of it and every finding resting on it from every report that carries one, regenerate those reports, add the domain to the do-not-quote list in 4.5 and the do-not-fetch list in 6.4, and tell the publisher what we did and when. Third, until the response-header and ai.txt checks are built, robots instructions and clause 6.4 are the mechanisms that work in practice, and building the other two is a condition of the first report being produced. This clause is reviewed at every release of this notice.

6.3 A reserved source is named, not hidden. Where a publisher has excluded us, the report will say the source was unavailable rather than leaving a silent gap, so a reader sees that the record has a hole instead of assuming there was nothing there.

6.4 Asking us to stop, and the do-not-fetch list. Write to [email protected] naming the domain, and tell us what your connection to it is. A domain we are asked to stop reading goes onto a do-not-fetch list we keep from the effective date of this notice, alongside the do-not-quote list in 4.5. We will not ask you to justify the request and we will not ask for a reason.

We do need to be satisfied that the request comes from someone who speaks for the domain, because an exclusion asked for by anyone else is a stranger's instruction about somebody else's site. A message sent from an address at that domain, or an exclusion of our fetcher in that domain's own robots or ai.txt file, satisfies that by itself. Otherwise we will ask you once for something showing your authority, and a clear answer is enough. We add the domain to both lists within one business day of being satisfied, and we confirm it to you.

The exclusion applies to material collected from that point on. A report already in production is completed on the sources it had, and we do not unpick reports already delivered. A source we cannot read because its publisher has excluded us is named as unavailable under 6.3; that is a hole in the record available to us rather than a failure to deliver what we described, clause 9 of Report Accuracy and Public Data governs source volume, and nothing in this clause affects a right a buyer has by law.

6.5 Our own reservation, stated in one place. We make no reservation against search engines and AI answer engines crawling, indexing, quoting or summarising the public marketing and legal pages of this site. That is deliberate: we publish that material to be found, and being read by an answer engine is the point of it.

Two things sit outside that permission, and they are stated as reservations rather than as conditions, because a condition is only as good as the willingness of the party reading it. We reserve against reproducing any figure from a published sample report - a score, a dimension value, a band - separated from the marking that names it synthetic; clause 12.3 states that reservation in full, and it is the only reservation we make about the samples, which in every other respect are offered for crawling, indexing, quotation, summarisation and training like every other public page of this site. And we reserve against bulk copying of this site, or any substantial part of it, for republication, for resale, or to build a competing product, dataset or model offering. Delivered reports are private documents outside every permission on this site, as clause 12.4 states. Section 12 gives effect to these reservations and governs automated access to this site.

An earlier text of this clause reserved the samples in full and reserved every page of this site against training, while section 12 said the opposite of both. The change record below dates the correction; section 12 was the position, and this clause now says so.

6.6 Retention as evidence is a separate act. We keep what we read so a buyer can check a finding against its source and so we can answer a complaint about one. That is a delivery and evidencing purpose rather than a mining purpose, and we treat it as needing its own answer. What is kept, and for how long, is in the Data Retention notice.

7. Databases, and review platforms specifically

7.1 Our own material. We claim whatever database right subsists in the collection of sample reports published on this site, and no more than subsists. We claim no database right, and no other right in the nature of a property right, in any collection of delivered reports or of the evidence behind them, and we do not build one: report content and the evidence behind it are destroyed on the schedule in the Data Retention notice, that material is never accumulated into a corpus, a feed or a searchable archive, and clause 2.5 of that notice is the authoritative statement of what it may be used for. Clause 6.2 of AI Transparency and Limitations records the same thing facing the other way: we build no accumulating record about a person who is not our customer.

7.2 Other people's databases. A database right is separate from copyright. It protects the investment its maker put into obtaining, verifying or presenting the contents, it can be infringed by taking a substantial part or by repeated and systematic taking of insubstantial parts, and it does not require that any protected expression was copied at all. We treat it as a live constraint on how we read aggregated sources rather than as a footnote to copyright.

7.3 Review platforms. Where a report describes what customers say about a company, we read public review pages. What a report carries is derived output and nothing else: the platform name, the number of reviews counted, the period they cover, the distribution of ratings, and the recurring themes with their counts. Any quoted phrase is subject to the caps in section 4. No more than three review platforms will be read for any one report, and the report will print, for each platform, how many review pages were read and how many reviews were counted, so that this clause can be checked against the report rather than taken on trust.

Retention is a separate question from what a report carries, and we answer it here rather than leave it to be worked out from two documents. A page we fetch, including a review page, is held as evidence on the schedule in the Data Retention notice, for the purpose stated in 6.6 and for no other. What we do not do, and will not do: hold a review record beyond that schedule; hold it for any purpose other than evidencing the report it belongs to; reuse it in a later report; make it available to anyone as a corpus, dataset, feed or searchable archive; or license, sell or supply it onward. The number of review records read per platform per report is capped. We do not state the figure here today, because a number published ahead of the control that enforces it is a reassurance rather than a cap; the figure is published in this clause in the release that builds the check, and enforcing it by that check is a condition of the first report being produced.

7.4 The state of this section, as at the effective date above. The derived-only posture in 7.3 is a requirement we have adopted rather than a control we have observed running. Our internal specification has not yet been changed to match it, and the cap on review records per platform per report is not built. Both are conditions of the first sale. This clause is reviewed at every release of this notice and is removed in the release that makes it unnecessary, with the removal dated in the change record.

7.5 We hold no account with any platform. We do not register, do not sign in and accept no platform's terms, because we read only what a browser reaches without credentials. That is a fact about how we work, not an argument that a platform's rights do not reach us.

8. Our trademarks

8.1 The names we use as marks. LeMans Labs, Company X-Ray™ and Founder X-Ray™. Executive Package is a product name we use without a symbol. Intelligence Score and Personal Brand Score describe what they are and we claim no mark in them.

8.2 None of them is registered. We make no claim of registration anywhere. The ™ symbol marks a name we use as a trade mark; it is not a claim that a registry has granted anything. We do not use the ® symbol on any material, and if you ever see it on ours, it is an error and we would like to be told at [email protected].

8.3 If that changes. If a mark is registered we will name the registration and the office here, and the symbol will change then and not before. This section and the symbols on the website are maintained together, and they say the same thing today.

8.4 Using our names. You may name us and our products truthfully, for example to review us, report on us or say that you are a customer. That permission reaches our names and nothing else. It is not permission to publish a report or any part of one, which clause 8 of the Terms of Service does not allow, with or without attribution to us.

8.5 What you may not do with our names. Use them in your own product, service, company, domain or application name; use them in advertising in a way that suggests we are affiliated with you, endorse you, or have verified or certified anything about you; register them or a confusable variant as a mark or a domain; or use our logo without written permission.

8.6 No badge, and no score in advertising. Nobody may present a report, an Intelligence Score, a dimension value or a band as a rating, certification, award, accreditation or verification issued by us, whether about themselves or about anybody else. That includes badges and seals, comparative advertising, marketing or investor material, material aimed at a subject's customers or market, and any use of our name to lend weight to a statement about a third party. A score is an evaluation produced by an automated system against a published rubric, for one buyer's own decision. It is not a credential, we issue no credential of any kind, and we will not have one made out of it.

8.7 Who these restrictions reach. Sections 8, 9 and 12 restrict people who order a report from us and people who use this website. They do not restrict a person or organisation a report is about.

If a report names you and you did not buy it, you may say what you wish about it, about your score and about us. If a copy reaches you, you may publish it in whole or in part and give it to your advisers, your investors, a regulator, a court or the press. We assert no right of any kind against you for doing so, and 1.2 records that we do not claim copyright in the text of a report. What restrains a buyer is their contract with us, and it does not restrain you, because you are not party to it.

Where you are the buyer, one document governs and it is not this one. A report about you is yours: you bought it, it is your own record, and you may show it to anyone you judge fit, investors and a board included. A report about a company is not. It goes to your own people who need it for the decision it was bought for, and to your own professional advisers under a duty of confidence, and no further. Clause 8.3 of the Terms of Service states it in full - including one further limb for a buyer that is an investment firm - and governs; this notice claims no priority over it, and an earlier version of this clause did, which was wrong.

Clause 8.6 is a different thing from a restriction on what you may say, and it is not a contract term binding on you if you never agreed to one. It is our published position, and it is the ground on which we will object: presenting our score as a certification we did not issue misstates our name and misleads whoever it is shown to. Three things follow that we do rather than merely say. Every report will carry two legends on its face: a confidentiality legend, in the report header, on any shared view of it and on every page of a file exported from it, marking the document as supplied to one buyer and not for republication; and a legend stating that the score is not a rating, certification or verification and may not be presented as one. The file carries the confidentiality legend on every page rather than once, because a page of it can be photographed, forwarded or pasted on its own, and because the file is the copy we can no longer reach. The first is the artefact half of the restriction in clause 8 of the Terms of Service; the second is this clause. Where we learn that a score is being presented as a credential we will say publicly, and to whoever asks, that we issued no such thing. And a buyer who discloses a report to its subject under clause 8 of the Terms of Service does so on condition, binding on that buyer, that the buyer tells the subject in writing that 8.6 applies.

9. Other people's trademarks

9.1 They belong to their owners. Every company name, product name, mark and personal name appearing in a report or in a sample is the property of its owner. We use it to identify the subject, which is what those names are for and the only way to demonstrate the format on a recognisable subject.

9.2 No relationship is implied. Nothing on this site or in a report indicates affiliation, sponsorship, endorsement, partnership, certification or approval by any named party, and none should be inferred. Vendor names for the model families we query identify those models and nothing more.

9.3 Restraints we place on ourselves. Word marks in plain text only. No logo, no wordmark typeface, no brand colour scheme, no product imagery. No third-party mark in a page title as though it were a partnership, in paid search copy, or in our own name.

9.4 Samples. The published sample reports carry a disclosure that they are illustrative demonstrations of the format, that their measurements are synthetic and that they are not affiliated with or endorsed by any subject.

Anyone named in a sample, or anyone acting for them, may have that sample unpublished by writing to [email protected], which is monitored on every business day. We unpublish it within one business day of the message arriving, without argument, without asking for a reason and whatever we think of the merits, and we confirm to the sender once we have done it. Clause 10.5 of Report Accuracy and Public Data states the same rule in the same terms.

Unpublishing means the sample leaves this website and our sitemap, and that we ask any search engine or answer engine holding a copy to drop theirs. It does not mean we destroy our own record of it. We keep a dated, unaltered internal copy of every sample we have published, with a content hash and the dates between which it was live, for the period stated for order and dispute records in the Data Retention notice. We keep it because a buyer who bought while that sample was live was shown it, so it forms part of what we described to that buyer, and we have to be able to produce what we showed if that buyer, a card issuer or a consumer authority asks. That archive is not published, is not shown to anyone else and is used for nothing else.

9.5 If you think we have gone beyond identification. Section 10.

10. If you believe your rights have been infringed

10.1 What this route is for. Copyright, database rights, trade marks, and a licence you say we have exceeded. There is no account to create, no fee, and no particular form of words.

10.2 What this route is not for, because a faster one exists. If you are a natural person and your complaint is that something we have written or published is wrong about you, or that we have used your name, image or likeness, do not use this section. Clause 4 of Report Accuracy and Public Data and section 13 of If a Report Names You are faster and give you more: acknowledgement within one business day, suppression of the disputed statement while we examine it rather than after we have finished, and a decision on suppression within three business days.

You do not have to find the right address. A message to either address in 10.3, or to [email protected], reaches that route, and we move it there ourselves rather than answer it on this section's timetable. No timetable in this section applies to you, and nothing in this section reduces anything in those two documents.

10.3 Where to send it. [email protected], or by post to LeMans Labs OÜ, Valukoja 8/1, 11415 Tallinn, Estonia.

10.4 What helps us act quickly. Who you are and who you act for; the work, mark or right; where it appears, with a link or the address of the page; what you would like us to do; a statement that you believe in good faith the use is not authorised; and a statement that what you have told us is accurate. We ask for these because they let us act, not as conditions. A notice missing something still gets an answer.

10.5 What happens, and when. We acknowledge within two business days. We give you a substantive answer within fourteen days, which is the period clause 10 of the Terms of Service and the Refund Policy also state, so that one clock runs across our documents rather than three. If we need longer we tell you that, why, and the date by which you will have an answer, before the fourteen days are up. The clock runs from the day your message reaches us, not from the day we finish establishing who you are.

Two things do not wait on that clock. A complaint about a published sample is dealt with under 10.6 within one business day. A complaint that requires us to withdraw a delivered report is acted on as soon as the decision is taken, and what happens to the money is in the Refund Policy.

10.6 A sample is unpublished first. If your complaint concerns a published sample report, we take it down within one business day, pending review, whoever complains and whatever we think of the merits. That is a business decision rather than an admission: samples are replaceable and the dispute is not worth having. Clause 9.4 says what unpublishing means and what we keep.

10.7 A report is a different thing, and we will not pretend otherwise. A report is not published. It is delivered to one buyer. If a finding is wrong we can correct or remove it, mark it as contested and suppress it, and regenerate the document, as clause 7 of Report Accuracy and Public Data sets out. Our buyer is contractually required by clause 10 of the Terms of Service to stop using and circulating a superseded version and to withdraw copies already supplied. A correction reaches the report in the account and every link issued for it, because a link is read through us and the next request gets the corrected document. It does not reach a file exported before the correction: that file keeps the wrong sentence, we cannot recall it or overwrite it, and the only thing operating on it is the buyer's obligation to stop using a superseded version. The same is true of what a buyer transcribed while reading. We say so rather than implying we can reach either.

10.8 The outcome, either way, with reasons, and the record we keep of it. If we agree with you we remove or change the material and, where it makes sense, add the domain or the mark to a do-not-use list so the same thing does not recur. If we do not agree we tell you why.

We keep a register of every notice under sections 10 and 11 and of what happened to it: who wrote and on whose behalf, the date it arrived, what was complained of, the domain or mark concerned, what we decided, the date we acted, and the reference of any sample unpublished or any report corrected or withdrawn as a result. Where a notice leads to a report being withdrawn, the entry is linked to that order so the two can be produced together. That register, the do-not-quote list in 4.5, the do-not-fetch list in 6.4, the do-not-use list in this clause, and the record of a buyer refused under 11.3, are each listed as their own class, with its own retention period, in the Data Retention notice, which is the only place a period appears. We do not publish any of them, and we do not confirm to anyone else that a particular person, buyer or domain is on one. The register begins on the effective date of this notice and begins empty.

10.9 What we will not tell you. We will not identify the buyer of a report to anyone who asks us, and that limit applies to everyone equally, including to buyers asking about other buyers. Two things sit outside it. We will identify a buyer where a court, a regulator or the law requires it. And where a person a report names asks us, we will tell them what a report about them said, what sources it used, when it was produced, how long it lasted, every link issued for it – who it was issued to and when each view happened – and every time a file was exported from it. We will also tell them what those records mean, because the two halves are not alike: a link we ended, and we can say when; a file we did not end, because we cannot reach one. That is everything we hold about the use of that report other than the buyer's identity. If a Report Names You states the reasons for the limit and what a person named is entitled to instead.

10.10 If a report names you personally. If a Report Names You is faster and more specific than this section, and nothing in this notice or in any customer's contract limits the rights described there.

11. Counter-notice and repeat infringers

11.1 If we took something down and you say we were wrong. Write to the address in 10.3 with what was removed, why you say the removal was wrong, and how to reach you. We will consider it and answer within fourteen days, and we will say what we decided and why.

11.2 What we tell the other side. We will tell the original complainant that a counter-notice was received and what we decided. We will not pass on your contact details unless you ask us to.

11.3 Repeat infringement. Where a buyer repeatedly infringes third-party rights through this service, or repeatedly breaches the licence in clause 8 of the Terms of Service, we may revoke that licence, end online access to the reports in that buyer's account and refuse further orders from that buyer. Clause 10 of the Acceptable Use Policy sets out how enforcement works and what we do about a copy already in someone's hands.

A refusal of further orders is recorded against the buyer and applied before a payment is taken, so a refused buyer is told at the point of ordering rather than charged and reversed afterwards. That control does not exist today and building it is a condition of checkout opening.

This clause says nothing about money and creates no forfeiture. What happens to what a buyer has paid, in every case including this one, is stated only in the Refund Policy, which governs if this notice or any other page appears to say something different.

11.4 Notices we may decline to act on. We may decline to act on a notice only where it repeats a notice we have already answered and adds nothing new, or where it identifies no material, mark or right at all. If we decline, we say so and we say why.

We do not decline a notice on the ground that the material complained of is accurate, or that it is already public. Whether it is accurate is the question the notice asks us to examine, and it is not a reason to refuse to examine it. A complaint that a finding is wrong is dealt with on its evidence under clause 7 of Report Accuracy and Public Data, and never by whoever writes at greater length. Where the complainant is the person a statement is about, clause 4.2 of that notice applies from the moment the message reaches us, whatever route it came by and whatever we think of the merits.

11.5 What we publish, and what we do not host for anyone else. Everything published on this site is written by us or generated by our systems. It includes third-party material only as section 3 describes: source links, source titles, publication dates and short attributed quotations, with the rights in them staying with their owners under 3.4 and 9.1. We publish nothing supplied by a user. The only thing a customer submits is a subject name, a link and an optional line of context; none of it is published and none of it is shown to any recipient, as 2.2 states. Where the service offers a way to show a report to a recipient the buyer is permitted to choose, what is served is our own document, to that recipient, and it is not published, not indexed, not listed and not available to anyone else. We describe sections 10 and 11 as our own complaints procedure, and it is offered whoever you are and whatever the legal characterisation of what we do.

12. Automated access to this site

12.1 What this section governs. This section governs automated access to this website and to the published sample reports: who may crawl, index, ingest, quote or summarise them, and what we do not permit. Clause 9.4 of the Terms of Service and clause 7.6 of the Acceptable Use Policy point here for that question. Two related things are answered elsewhere and are not restated here: conduct against our systems and our technical controls, including circumventing a rate limit, is in clause 9 of the Acceptable Use Policy, and automated ordering is in clause 7 of that policy. Where this section and either of those appear to overlap, each governs only the subject named for it here.

12.2 What we permit. Search engines and AI answer engines may crawl and index the public marketing and legal pages of this site, and may quote a short extract from one of them alongside a link to it. Our robots file names the agents this is addressed to. We are not being tolerant; we publish that material to be found, and being read by an answer engine is the point of it.

12.3 The sample reports, and the one thing about them we do not license. The published sample reports are offered for crawling, indexing and citation like every other public page of this site: a search engine or an answer engine may link to a sample and show a short extract of it. What is not licensed, for samples specifically, is the reproduction of any figure from one - a score, a dimension value, a band - separated from the sentence beside it that marks the figure as synthetic. A sample is a demonstration of the format whose measurements are invented, as clause 10 of Report Accuracy and Public Data sets out, and a sample names a real subject: a number restated without its marking becomes an invented measurement of a named company presented as ours. Every such figure carries its marking on the page and in the page's structured data, which is how an automated reader is told, and this clause is how a deliberate one is answered.

A sample that scores, rates or bands a real, identifiable natural person is outside every permission in this section, is outside anything we are willing to publish, and is unpublished on request under 9.4. The published corpus meets that standard: the one sample about a person portrays an invented subject, says so at the top of its page and in its structured data, and scores nobody who exists. Keeping the corpus to that standard is a standing commitment of this notice, not a pending condition.

12.4 What we do not permit. Retaining any delivered report, or any part of one, in any system: delivered reports are private documents and are outside every permission in this section. Using a delivered report as training data. On the public site the position is the opposite and we state it plainly: the marketing pages and the sample corpus may be used as training data, because being known to the systems people ask is part of how this product is found, and a reservation we do not mean would be worth less than none. Reproducing an Intelligence Score, a dimension value, a band or a verdict sentence taken from a sample. Bulk copying, mirroring or systematic downloading of this site or any substantial part of it, whether for republication, for resale, or to build a competing product, dataset or model offering. Probing accounts, data or systems that are not yours, or circumventing a rate limit or a technical control. Automated ordering is a separate question and is governed by clause 7 of the Acceptable Use Policy.

12.5 The line between them, drawn by what is done rather than by who is doing it. What 12.2 permits is retrieval that produces a link and a short extract in an answer. What 12.4 forbids is the retention or training use of a delivered report, bulk copying of the site, and the restatement of a number we generated without its marking. We draw the line by the act rather than by the identity of the reader, because the same crawler performs both, and a rule that permits an actor rather than an act permits everything that actor does. Where a single act falls within both 12.2 and 12.4, clause 12.4 governs.

12.6 Reports are not public, and this is what actually keeps them out of an index. We do not publish, index or list reports. Reports and shared links now exist, so this clause says which of the protections it used to promise are in place and which are not, rather than leaving a future tense standing over a live feature.

What is in place is the access check. A shared report is served only to the person the link names, that person confirms the address the link was sent to with a code before the first view, every view is served by us, and nothing anonymous reaches the document at all – not a crawler, not a preview fetcher, not a platform unfurling a pasted address. That is the protection this clause now rests on.

What is not in place is the search-engine work: robots entries for the report and share routes, directives instructing search engines not to index or follow them, exclusion from our sitemap, and a link preview carrying our name and nothing else – no subject name, no score, no band. It is not built. It has a row in section 13 of the Security notice, and we name it rather than let a promise in the future tense read as a control. The reason it is worth building has not changed: a preview card is fetched and cached by whatever platform a link is pasted into, and a cached card survives revoking the link.

An exported file has no address at all, so nothing indexes it. That is not a protection and we will not present it as one. It means the file is outside anything we can do about it, which is clause 9.7 of the Security notice.

12.7 Our own fetcher. When the pipeline runs it will identify itself as LeMansLabsBot/1.0, with a URL in the user-agent string that resolves to a page describing the fetcher, the agent name to use to exclude us and the route in 6.4. It will obey the robots instructions set for that agent. Clause 6.4 is the faster route and does not require you to change any file.

12.8 We may block automated access at any time and without notice, including access this section otherwise permits, where it degrades the site for other people.

12.9 If our robots file and this section disagree. They are changed together, and a difference between them is our error rather than yours. Tell us at [email protected] and we will fix it. Until we do, two things apply. Our robots file is a technical instruction and not a licence, so nothing in it widens the permission granted by 12.2. And if you acted in good faith on either the robots file or this section as you found it, we will not treat what you did as a breach of this notice and we will bring no claim about it.

Change record

15 August 2026. Clause 6.5 was corrected to the position section 12 has stated since 9 August 2026: the published samples are offered for crawling, indexing, quotation, summarisation and training, and the one reservation about them is the figure-marking rule in 12.3. The two clauses had said opposite things since the 9 August pass rewrote section 12 and left 6.5 standing; 12.9 forgives a disagreement between the robots file and this text, not between two clauses of it. The condition in 12.3 is recorded as satisfied - the founder sample's subject is invented and labelled as such on the page and in its structured data - and the clause now says "real, identifiable natural person", which is what it was always for. Clause 12.10 is removed under its own instruction: the robots file, the page directives and the sitemap now match the reservations as stated. The robots file does carry exclusions for the API, the signed-in area, share links and the monitoring route; those are access boundaries for surfaces that were never offered, not reservations of published content, and 12.2 remains the statement of what is offered.


This notice is part of the LeMans Labs legal set. Read it with the Terms of Service, which is the contract, carries the licence in clause 8, and identifies the company you are dealing with and the address for notices; the Acceptable Use Policy, which governs conduct and ordering; Report Accuracy and Public Data and AI Transparency and Limitations, on what a report is, what it quotes and how it is produced; the Refund Policy, which is the only statement of what happens to the money; the Data Retention notice, on how long we keep what we read; the Privacy Policy, which names the controller and the contact for data protection questions; Sub-processors, which names every recipient; and, if a report names you, If a Report Names You. There is no separate imprint page. Provider identification and the address for notices are in the Terms of Service, and the controller and the contact for data protection questions are in the Privacy Policy.