Terms of Use

001 / Scope

What this covers, and what it deliberately does not.

This page is a set of website terms. It applies to saptiva.com: the pages, the copy, the images, the get started form, and the published inboxes. If you are reading the site, you are inside it.

It is not a platform agreement, and it is not written to look like one. There is no plan to pick, no account to open, and no button here that binds you to a service. That is a deliberate consequence of how Saptiva AI sells: an engineer talks to your team, a deployment is scoped, and a contract is negotiated. The commitments that matter to you are made there, in writing, by people who read them.

Where this page and your agreement say different things, your agreement governs. It was negotiated. This was published.

002 / Who we are

Saptiva AI.

Saptiva AI is an artificial intelligence infrastructure company headquartered in Mexico City. We build the control plane enterprises run their AI on: Saptiva Studio for the applications, and frIdA for routing workloads across environments and writing a signed record of every decision.

Throughout this page, "we" and "us" mean Saptiva AI, and "you" means the person reading the site, or the organization on whose behalf you are reading it.

003 / Using this site

Read it, quote it, send it to your risk committee.

This site exists to be evaluated. Read it, print it, quote it, link to it, forward it internally, and put it in front of the people at your organization who will have opinions about it. None of that needs our permission, and asking for permission to link to a public web page has always been a strange convention.

What we ask you not to do

• Do not attempt to breach, disrupt, or degrade the site or the infrastructure serving it, or to reach any part of it you were not offered.
• Do not extract the site in bulk by automated means at a volume that affects its availability for anyone else. Ordinary crawling and archiving are fine.
• Do not present yourself as Saptiva AI, as speaking for us, or as endorsed by us when you are not.
• Do not use the site to do something unlawful, or to send us content you have no right to send.
• Do not use our brand marks to imply a partnership, a certification, or a customer relationship that does not exist.

If we ever have to restrict someone's access to this site, it will be for one of the reasons above and nothing else. We do not moderate opinions about us.

004 / Our material

What is ours, and what you may do with it anyway.

The copy, design, diagrams, product imagery, and software on this site belong to Saptiva AI or to the people we licensed them from. The Saptiva AI name, the wordmark, frIdA, Saptiva Studio, Savant, and KAL are our marks. Third-party marks that appear on the site, including those of our partners and customers, belong to their owners and appear with their permission or as accurate factual reference.

Journalists, analysts, and anyone reporting on us

You have more latitude than this section might suggest, and it is granted on purpose. The Press page licenses our logo pack, brand guidelines, executive photography, and approved product imagery for editorial use with attribution, and states plainly that everything on it is fair game for accurate reporting and analysis. Where this section and the press page differ, the press page governs for editorial use.

Quoting this site for reporting, analysis, commentary, teaching, or a competitive evaluation is fair and needs no permission. We would rather be quoted accurately than not quoted, and if we have got something wrong, press@saptiva.com takes corrections.

What is not granted

Nothing here gives you a license to republish the site as your own, to recolor or redraw the wordmark, or to use our marks in a way that suggests we endorse you.

005 / What you send us

We claim nothing you send us.

When you fill in the get started form or write to one of our inboxes, you keep every right you had in what you sent. We take no ownership, and we take no license. Not a perpetual one, not an irrevocable one, not a sublicensable one, not one that survives anything. There is no clause elsewhere in this document that quietly restores one.

What we do with it is what you would expect: an engineer reads it and replies, we use it to understand and scope what you need, and we keep it while the conversation is useful. That is the whole permission we are asking for, and section 003 of the Privacy Policy describes exactly what happens to it.

Please do not send confidential material through the form

A web form is the wrong container for anything sensitive, and we would rather say so than accept it and hope. Describe the shape of the problem. If getting further requires real detail about your systems, your data, or your customers, tell us and we will sign a mutual NDA first, and then have that conversation properly. We already do this routinely for security and compliance reviews.

Ideas and feedback, honestly

We do not run an idea submission program, and unsolicited product ideas put both of us in an awkward position: we may already be building the thing you are describing, and we cannot treat what arrives unsolicited in a public inbox as confidential. So if you send us a product idea without an agreement in place, we will read it, we may already be working on it, and we are not going to pretend otherwise. If you tell us something is broken and we fix it, we do not owe you a royalty and we are not claiming anything of yours either.

006 / The platform

Nothing on this website is a commitment to you.

The product pages describe what the platform does and what it has done. Deployment timelines, environment modes, control descriptions, capability tables, and customer outcomes are descriptions, current when published and maintained in good faith. They are not an offer, not a warranty, not a service level, and not a guarantee of a result in your environment.

This is the same thing the Security & Compliance page says about itself, and it is worth repeating here: the claim on a page is not the commitment. The commitment is in the agreement. Which controls are engaged, what residency is guaranteed, what happens when something breaks, and what we are liable for are all decided in your Master Services Agreement and Data Processing Agreement, against the specific deployment proposed for your environment.

Two things this does not water down. The response times published on the Contact page are how we actually operate, and we intend to be held to them as a matter of conduct, though they are not contractual service levels until an agreement makes them so. And the security posture on the Security & Compliance page is a maintained, versioned document that we correct when it stops being true, including the parts that say what we do not yet have.

To see the agreement set before you are in a process, write to legal@saptiva.com. We share it with any party in an active commercial or procurement evaluation.

007 / Documentation

Developer documentation.

Our developer documentation is published on an external service and linked from the footer of every page. It is reference material: it describes how the platform behaves, and it changes as the platform does. It is not a contract, it does not grant access to anything, and the service hosting it has its own terms and its own privacy practices, which are not ours.

Access to the platform itself, including any API credential, comes from an agreement, not from reading the documentation.

008 / Third parties

Where this site points elsewhere.

This site links to our documentation, to partner and customer material, and to public sources we cite. It loads typefaces from an external font service, which is disclosed in section 003 of the Privacy Policy because it has a privacy consequence.

We do not control those destinations, we do not adopt their content, and their terms and privacy practices are their own. When we cite a public source, we cite it because we think it is accurate and relevant, not because we have any relationship with whoever published it.

009 / No warranty

About this website, specifically.

This website is provided as it is. We do not warrant that it will be available without interruption, that it will be free of error, or that every figure on it is current at the moment you read it. We correct what we find and what you tell us about, and pages that carry a version and review date carry them so you can see how fresh they are.

Read this paragraph narrowly. It is about a website. It says nothing about the platform, which comes with the warranties your agreement gives it.

010 / Liability

Scoped to the website, and to nothing else.

To the extent the law allows, Saptiva AI is not liable for loss arising from your use of this website or from reliance on material published here, including a decision you made on the strength of a page rather than on the strength of an agreement.

Nothing here excludes liability that the law does not permit us to exclude.

011 / Security research

Good-faith research is welcome.

If you find a security issue in this website or in a Saptiva AI platform surface, report it to security@saptiva.com. That inbox goes to our security lead and is monitored continuously.

We will not pursue or support a claim against a researcher who acts in good faith. Good faith means: you report promptly to that address, you give us a reasonable window to fix the issue before you disclose it, you do not access, alter, destroy, or exfiltrate data that is not yours, you do not degrade the service for anyone else, and you stop and tell us as soon as you have proved the issue exists. Testing that ignores those lines is not research, and this paragraph does not cover it.

The Security & Compliance page carries our full posture, including incident notification and what we do after one.

012 / Changes and contact

Versioned, and answered by a person.

When these terms change, the version number and the review date at the top change with them. We do not treat your continued reading of a public web page as acceptance of a new version, and we do not reserve a right to change them silently.

Questions about this page, or a request for the current agreement set, go to legal@saptiva.com. Personal data questions go to privacy@saptiva.com, and the Privacy Policy covers them in detail. Security goes to security@saptiva.com. Every one of those is read by a person on the team that can act on it, which is the argument the Contact page makes at greater length.