Terms of Service

Version 2026-08-12. Effective 2026-08-12.

These terms are a contract between you and Andrii Sukhanov, a natural person trading as an independent professional (autónomo) registered in Spain, NIE Z2338955K, using merkleset as a trading name ("we", "us"). There is no company: your counterparty is an individual established in Spain. Full identification details are in the Legal Notice.

These terms govern your use of the merkleset website, our API, and the datasets we deliver. Read them before you create an account. If you do not agree, do not use the service.

We sell only to businesses, professionals and other organisations. The service is not offered to consumers, and by accepting these terms you confirm you are acting for purposes related to your trade, business, craft or profession.

1. Definitions

1.1 Service — the merkleset website, the public API, the account area, and the delivery of dataset releases.

1.2 Dataset — a named corpus in our catalogue, such as the US federal procurement corpus or the US federal regulatory corpus.

1.3 Release — one published, versioned artifact set for a dataset. A release contains JSONL data files and a manifest that holds a Merkle root over the hashes of the data lines, per-file hashes and byte counts, and quality metrics.

1.4 Provenance fields — the per-record fields we publish with every record: source_url, fetched_at, http_status, raw_sha256, extracted_sha256, snapshot_ref, pipeline_version, parser_id.

1.5 Subscription — your right, during a paid term, to receive the releases we publish for the datasets in your plan and to download them.

1.6 Data Licence — the Data License Agreement published alongside these terms. It governs what you may do with delivered data. Where these terms and the Data Licence differ about permitted use of data, the Data Licence controls.

1.7 Documentation — the dataset descriptions, schema notes and release notes we publish on the website.

2. Your account

2.1 You must be at least 18 years old and acting for an organisation, a business or a profession.

2.2 The person who accepts these terms confirms they are authorised to bind that organisation.

2.3 Login is passwordless. We email a one-time code to your address; entering it issues access and refresh tokens to your browser. You are responsible for the security of the mailbox that receives those codes and for activity under your account.

2.4 You must give us a working email address and keep it current. Notices we send to that address are effective when sent.

2.5 Do not share your account, your tokens, or your download links outside the scope your plan covers. Sharing rules are in the Acceptable Use Policy.

3. Formation of a subscription

3.1 Plans, their contents and their prices are published on our pricing page. We do not restate prices here so that one page stays the single source of truth.

3.2 A subscription is formed when we confirm your order in writing, including by email. A plan description on the website is an invitation to order, not an offer we are bound to accept.

3.2.1 How an order is placed today. You tell us which plan and which datasets you want — the contact form is the simplest route — we agree the order, and the payment transaction is processed and invoiced by Stripe as merchant of record (clause 4.3).

3.2.2 Access is enabled once the order is confirmed. Provisioning is a deliberate step on our side, not an automatic consequence of a payment event, so do not plan for access to appear in the same instant a payment clears. We enable it promptly and confirm by email.

3.3 Free and demo access is a subscription for the purposes of these terms, except that no fees are payable and either side may end it at any time.

3.4 Enterprise plans are sold under a separate written agreement. Where such an agreement exists and conflicts with these terms, that agreement controls.

4. Fees, taxes, renewal and price changes

4.1 Fees are stated in the currency shown on the pricing page and are exclusive of taxes.

4.2 Unless your order says otherwise, subscriptions are monthly and renew automatically at the end of each period until cancelled.

4.3 Two layers, and where the line sits. Payment is due in advance of each period.

1. We provide the service and grant the Data Licence. Your agreement about the data, the releases and their use is with us.
2. Stripe processes the purchase transaction as merchant of record. Stripe invoices you, takes payment, determines, collects and remits any applicable taxes, and handles payment disputes and chargebacks.

So: a question about the data, a dataset, a release or your entitlement comes to us at contact@merkleset.com. A question about a charge, an invoice, a refund execution or a card dispute is handled through the merchant of record's flow, and we will point you to it. We never see or store your full card details.

4.4 Taxes. Prices on the pricing page are quoted exclusive of applicable taxes for business customers. As merchant of record, Stripe determines which taxes apply from your location and status, adds them to the amount payable, and remits them to the relevant authorities. Any tax identification number needed for your jurisdiction's treatment is supplied by you as part of the order and payment process. Your invoice for the transaction comes from the merchant of record. Nothing in this clause is tax advice; take your own.

4.5 We may change prices. A change takes effect for you at the start of your next period and only if we gave you at least 30 days' notice by email. If you do not accept the new price, cancel before that period starts.

4.6 If payment fails, is reversed, or an invoice is more than 14 days overdue, we may suspend delivery and downloads for the affected datasets after giving you notice. Suspension does not end your licence to data already delivered.

4.7 Cancellation and refunds are governed by the Refund and Cancellation Policy.

5. What we deliver, and what we do not promise

5.1 During a paid term we publish releases for your datasets at the update cadence stated on the dataset page for your plan.

5.2 Each release covers the collection window it was produced for. A release is not a cumulative snapshot of the whole corpus. To reconstruct a full corpus you need each window's release. Record identifiers and chunk identifiers are stable across releases, so an unchanged record appears in two overlapping windows with identical identifiers and hashes.

5.3 Downloads are delivered as time-limited signed URLs, issued only to an account that holds an active entitlement for that dataset. Signed URLs expire; request a new one when they do.

5.4 Your right to download ends when your subscription ends. Your licence to the data you already received does not. That two-part rule is set out in clauses 3.1 to 3.4 of the Data Licence. Because access ends, keep your own copies of any release you rely on. We are not obliged to keep any release available for download after your subscription ends, and we do not act as your archive.

5.5 We do not commit to an uptime level, a recovery time, or a support response time under these terms. We use commercially reasonable efforts to keep the service available and to answer questions about the data and the service sent to contact@merkleset.com. Payment-related support — invoices, charges, refund execution, disputes — is handled by the merchant of record under clause 4.3. Any service level commitment must be in a separate signed agreement to bind us.

Two operational facts you are entitled to weigh before you buy: the service runs on a single host, and there are today no automated backups and no automated monitoring. That is why this clause promises effort rather than a number.

5.6 We may change how a dataset is produced, which upstream endpoint we read, or how records are chunked, provided we do not reduce the substance of what the dataset delivers during your current period. Material changes are described in the documentation.

5.7 We may retire a dataset. If we retire a dataset you subscribe to, we will give you at least 30 days' notice and refund the unused part of any fee you paid in advance for it.

5.8 Data quality is measured and published, not promised. Release manifests state completeness figures for that release. Some source records carry no body text; we publish them with metadata and count them in the completeness figure rather than hiding them.

6. Source material, and how we treat personal data in datasets

6.1 Our shipping datasets are built from official public United States federal government sources. Works of the US government are not subject to copyright in the United States.

6.2 Our collection practice, which we apply to every source: we collect logged out; we do not create or use accounts on source sites; we honour robots.txt and machine-readable text-and-data-mining reservations; we keep request rates modest; and we stop collecting from a source on receipt of a credible demand to stop.

6.3 Datasets are built to contain no personal data. Screening runs inside the pipeline before any published hash is computed, so the hashes we publish describe exactly the bytes we deliver. Where a source has a column boundary that lets us avoid reading contact fields at all, we do not read them.

6.4 Screening is systematic, and its scope is deliberately narrow and documented. It is not a warranty that no personal data can appear in a dataset. In particular, a person named in running prose without a contact label — for example an official named in the middle of a regulatory preamble — is not removed. The published scope and its known gaps are described in the documentation and in clause 7 of the Data Licence. Clause 7.5 of the Data Licence and sections 1 and 2 of the Acceptable Use Policy state what you must not do with any residual personal data.

7. Your obligations

7.1 Use the service and the data only as the Data Licence and the Acceptable Use Policy allow.

7.2 Do not attempt to re-identify any individual from any data we deliver.

7.3 Do not use the data to make decisions about an individual's eligibility for employment, credit, insurance, housing or a similar benefit.

7.4 Do not probe, load-test or circumvent our authentication, rate limits or entitlement checks, and do not automate our delivery endpoints beyond what the published API and rate limits allow.

7.5 Comply with the export control and sanctions laws that apply to you.

7.6 If you receive a takedown demand, a data subject request or a regulator's request that concerns data we supplied, tell us at legal@merkleset.com promptly so we can respond in the same way.

8. Intellectual property

8.1 We own the service: the software, the pipeline, the schema, the chunking, our derived and enriched fields, the provenance layer, the manifests, the compilation of records into a dataset, and our name and marks. Nothing in these terms transfers that ownership.

8.2 The underlying US federal source material is a government work and we claim no copyright in it. What we license to you is our collection, cleaning, normalisation, chunking, enrichment, provenance layer and compilation. See the Data Licence.

8.3 You own your own data, your models, your embeddings and anything else you build. We claim no rights in them.

8.4 If you send us feedback, we may use it without obligation to you. Do not send us anything confidential as feedback.

9. Warranties and disclaimer

9.1 We warrant that we will provide the service with reasonable skill and care, and that we collected the source material in the way clause 6.2 describes.

9.2 EXCEPT AS STATED IN CLAUSE 9.1, THE SERVICE AND THE DATA ARE PROVIDED "AS IS" AND "AS AVAILABLE". WE MAKE NO OTHER WARRANTY, EXPRESS OR IMPLIED, AND WE SPECIFICALLY DISCLAIM ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, ACCURACY, COMPLETENESS, AND NON-INFRINGEMENT OF THIRD-PARTY CONTENT INSIDE THE SOURCE MATERIAL. WE DO NOT WARRANT THAT THE SERVICE WILL BE UNINTERRUPTED OR ERROR FREE, THAT ANY DATASET IS COMPLETE OR FREE OF PERSONAL DATA, OR THAT ANY RESULT YOU OBTAIN FROM A MODEL BUILT ON THE DATA WILL BE CORRECT.

9.3 The data describes public records. It is not legal, financial, procurement or compliance advice, and it must not be the only basis for a decision that carries legal or financial consequence.

10. Limitation of liability

10.1 NEITHER PARTY IS LIABLE FOR INDIRECT OR CONSEQUENTIAL LOSS, OR FOR LOST PROFITS, LOST REVENUE, LOST DATA, LOSS OF BUSINESS OPPORTUNITY, OR THE COST OF RE-TRAINING OR RE-EMBEDDING A MODEL, EVEN IF THAT LOSS WAS FORESEEABLE.

10.2 OUR TOTAL LIABILITY FOR ALL CLAIMS ARISING OUT OF OR RELATING TO THESE TERMS, THE SERVICE OR THE DATA IS LIMITED TO THE FEES YOU PAID US IN THE 12 MONTHS BEFORE THE EVENT GIVING RISE TO THE FIRST CLAIM. IF YOU HAVE PAID NO FEES, OUR TOTAL LIABILITY IS LIMITED TO {{LIABILITY_CAP_FLOOR}}.

10.3 The limits in clauses 10.1 and 10.2 do not apply to: your obligation to pay fees; our indemnity under clause 11.1; liability for death or personal injury caused by negligence; or fraud, wilful misconduct or gross negligence. They apply only to the extent Spanish law allows a limitation of liability, and nothing in this clause excludes or limits liability that cannot lawfully be excluded or limited.

10.4 You must bring any claim under these terms within 12 months of the date you first knew, or reasonably should have known, of the facts giving rise to it, except where a longer period is required by law.

11. Indemnities

11.1 Our indemnity. On the Pro and Enterprise plans only, we will defend you against a third-party claim that our collection of the source material, or the compilation and delivery of a dataset to you, infringed that third party's intellectual property rights or breached its terms of access, and we will pay damages finally awarded or agreed in settlement. This indemnity is subject to clause 10.2 and applies only while your subscription fees are paid up. It is not offered on the free or Starter plans, which are provided as described in clause 10.

11.2 Your indemnity. You will defend us against a third-party claim arising from your use of the data outside the Data Licence, your breach of the Acceptable Use Policy, or your use of the data to make decisions about individuals, and you will pay damages finally awarded or agreed in settlement.

11.3 Conditions. The indemnified party must notify the other promptly, give the indemnifying party control of the defence, and cooperate at the indemnifying party's cost. No settlement that admits fault or imposes an obligation on the indemnified party may be made without its consent.

11.4 Our indemnity does not cover a claim caused by your modification of the data, your combination of the data with other material, or your continued use after we asked you to stop using a specific record or release.

12. Suspension and termination

12.1 You may cancel at any time under the Refund and Cancellation Policy.

12.2 We may suspend your access immediately if we reasonably believe you are breaching clause 7, the Acceptable Use Policy or the Data Licence, if your payment is overdue after notice, or if suspension is needed to protect the service or comply with the law. We will tell you why and, where the cause is curable, give you a chance to cure it.

12.3 Either party may terminate for material breach that is not cured within 30 days of written notice.

12.4 We may terminate a free or demo subscription at any time.

12.5 On termination: your right to download and to receive new releases ends; your licence to releases already delivered survives as the Data Licence provides; each party's accrued payment obligations survive; and clauses 8, 9, 10, 11, 14 and 15 survive.

13. Changes to these terms

13.1 These terms are versioned by effective date. The version at the top of this document identifies its exact text.

13.2 We record which version you accepted, when you accepted it, and a SHA-256 hash of the exact text you were shown. That record is how we identify what you agreed to, rather than pointing at a page that may since have changed.

13.3 When we publish a new version, we ask you to accept it the next time you sign in. If you do not accept it, you may keep using data already delivered under the version you did accept, but we may decline to deliver new releases until you accept.

13.4 A new version never changes what you agreed to in the past. It applies from the date you accept it.

14. Notices

14.1 Notices to you go to the email address on your account.

14.2 Notices to us go to legal@merkleset.com and, if the notice starts a legal proceeding or a formal dispute, also by post to Andrii Sukhanov, Av. Marítima 2, puerta 05c, {{POSTAL_CODE}} Los Silos, Santa Cruz de Tenerife, Spain.

14.3 Email notices are effective on the day they are sent, unless the sender receives a delivery failure.

15. Governing law and venue

15.1 These terms are governed by the laws of the Kingdom of Spain.

15.2 The courts of Santa Cruz de Tenerife (Canary Islands, Spain) have exclusive jurisdiction over any dispute, and each party submits to that jurisdiction. Either party may still seek urgent injunctive relief in any court with jurisdiction to protect its intellectual property or confidential information.

15.3 The United Nations Convention on Contracts for the International Sale of Goods does not apply.

15.4 Before starting proceedings, each party will try in good faith for 30 days to resolve the dispute by discussion, starting with a written notice to the other party's notice address.

16. General

16.1 Assignment. You may not assign these terms without our written consent, except to a successor of your business that is not our competitor. We may assign them to a successor of our business, including on incorporation of the business into a company.

16.2 Entire agreement. These terms, the Data Licence, the Acceptable Use Policy, the Refund and Cancellation Policy, the Privacy Policy and your order form are the whole agreement about their subject matter, and replace any earlier discussion or proposal.

16.3 Severability. If a clause is unenforceable, the rest stays in force and the clause is read down to the extent needed to make it enforceable.

16.4 No waiver. A delay in enforcing a right is not a waiver of it.

16.5 Force majeure. Neither party is liable for a failure caused by an event outside its reasonable control, other than a failure to pay.

16.6 Independent parties. Nothing here creates a partnership, agency or employment relationship.

16.7 Publicity. We will not use your name or logo as a customer reference without your written consent.

16.8 Language. English is the authoritative language of these terms. We may publish translations for convenience; if a translation differs from the English text, the English text controls, and the English text is the one whose hash is recorded with your acceptance.

17. Contact

Questions about these terms: legal@merkleset.com. Commercial and support questions: contact@merkleset.com.