Skip to content
in/guard/out
in/guard/out

Legal

Terms of Service

Last updated: July 5, 2026

The plain-English summary: we build the Service carefully and operate it in good faith, but guardrails reduce risk rather than eliminate it – you stay responsible for your applications and their actions, and our liability is limited accordingly. The details, which control, are below.

1. Who we are and what these terms cover

These Terms of Service (the “Terms”) govern your access to and use of the in/guard/out website, the hosted demo, the guardrails proxy service, its dashboard, APIs, and related software (together, the “Service”), operated by in/guard/out (“we”, “us”, “our”). By creating an account, requesting demo access, or using the Service, you agree to these Terms on your own behalf and on behalf of the organization you represent. If you do not agree, do not use the Service.

We are an early-stage company and the Service is under active development. Parts of it are offered as a limited demo, beta, or early-access functionality, and may change, be interrupted, or be discontinued at any time.

2. The Service

The Service is a guardrails layer that sits between your applications and third-party or self-hosted language models. It applies configurable checks to requests, responses, and agent tool calls – such as PII screening, grounding and fact-checking, format enforcement, content safety, and agent policy enforcement – and provides logging, metering, and a dashboard.

You are responsible for your applications, your prompts, your end users, the models you route traffic to, and the actions your systems take. The Service adds checks around your traffic; it does not operate your applications for you.

3. The nature of guardrails: no guarantee of detection

This section is important. Guardrails – ours and anyone’s – are risk-reduction measures, not certainties. Some of the Service’s checks are deterministic; many depend on statistical models, third-party classifiers, heuristics, and configuration you control. By design and by the nature of the problem, they can produce false negatives (failing to detect a hallucination, an injection, a data leak, an unsafe action, or other harmful content or behavior) and false positives (flagging or blocking legitimate content or actions).

Accordingly, while we work hard to make the Service effective, you acknowledge and agree that: the Service is not guaranteed to detect, block, repair, or prevent any particular error, attack, leak, action, or harm; a passed check is not a representation that content is accurate, safe, lawful, or fit for reliance; and the Service is not a substitute for human review, testing, monitoring, or your own legal and compliance processes. You remain solely responsible for consequential decisions and actions taken by or through your applications, including those an AI agent takes with your credentials.

4. Accounts and API keys

You must provide accurate information when signing up and keep your credentials and API keys confidential. You are responsible for all activity under your account and keys. Notify us promptly at the contact address below if you suspect unauthorized use. We may suspend accounts that we reasonably believe are compromised or in breach of these Terms.

5. Acceptable use

You agree not to use the Service to:

  • violate any law or regulation, or infringe anyone’s rights;
  • send content or traffic you do not have the right to process, including personal data you lack a lawful basis to process;
  • probe, disrupt, overload, or circumvent the Service’s security or metering, except for security testing we have authorized in writing;
  • build or train a competing service using non-public elements of the Service; or
  • resell or provide the Service to third parties as a stand-alone offering without our written agreement.

6. Your content and data

You retain all rights to the prompts, completions, documents, tool traffic, policies, and other material you submit to the Service (“Customer Content”). You grant us the limited rights needed to operate the Service: to process, transmit, transform (for example, screening PII into placeholders), store, and display Customer Content to you, and to route it to the upstream model providers and tools you configure.

We may use aggregated, de-identified usage data (such as request counts, stage timings, and violation rates) to operate and improve the Service. We do not use Customer Content to train models, and we do not sell it.

How we handle personal data is described in our Privacy Policy.

7. Third-party providers and models

The Service routes your traffic to language-model providers, search services, and other upstreams that you select and configure. Your use of those services is governed by their terms, and their availability, pricing, behavior, and output are outside our control. We are not responsible for what a third-party model produces, for its errors or downtime, or for changes those providers make. Where you supply your own provider API keys, you are responsible for those accounts and the charges they incur.

8. Plans, fees, and changes

Paid plans, allowances, and pricing are described on our pricing page or in an order we agree with you. Fees are exclusive of taxes and, except where required by law or expressly stated otherwise, non-refundable. We may change prices and plan structures with reasonable notice, effective for subsequent billing periods. Demo and early-access use may be free, rate-limited, and withdrawn at any time.

9. Disclaimer of warranties

THE SERVICE IS PROVIDED “AS IS” AND “AS AVAILABLE”, WITH ALL FAULTS AND WITHOUT WARRANTY OF ANY KIND. TO THE MAXIMUM EXTENT PERMITTED BY LAW, WE DISCLAIM ALL WARRANTIES, EXPRESS, IMPLIED, OR STATUTORY, INCLUDING WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, ACCURACY, AND UNINTERRUPTED OR ERROR-FREE OPERATION. WITHOUT LIMITING THE FOREGOING, WE DO NOT WARRANT THAT THE SERVICE WILL DETECT, BLOCK, REPAIR, OR PREVENT ANY PARTICULAR CONTENT, ERROR, ATTACK, DATA LEAK, OR ACTION, OR THAT CONTENT PASSING THROUGH THE SERVICE IS ACCURATE, SAFE, OR LAWFUL.

10. Limitation of liability

TO THE MAXIMUM EXTENT PERMITTED BY LAW: (A) WE WILL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, OR FOR ANY LOSS OF PROFITS, REVENUE, DATA, GOODWILL, OR BUSINESS, EVEN IF ADVISED OF THE POSSIBILITY; (B) WE WILL NOT BE LIABLE FOR DAMAGES ARISING FROM CONTENT, DECISIONS, OR ACTIONS OF YOUR APPLICATIONS, YOUR AGENTS, YOUR END USERS, OR THIRD-PARTY MODELS AND SERVICES, INCLUDING HARMS THE SERVICE DID NOT DETECT OR PREVENT; AND (C) OUR TOTAL AGGREGATE LIABILITY ARISING OUT OF OR RELATING TO THE SERVICE OR THESE TERMS WILL NOT EXCEED THE AMOUNTS YOU PAID US FOR THE SERVICE IN THE THREE (3) MONTHS BEFORE THE EVENT GIVING RISE TO THE CLAIM, OR ONE HUNDRED U.S. DOLLARS (US$100) IF YOU HAVE PAID US NOTHING.

Some jurisdictions do not allow certain exclusions or limitations; in those jurisdictions, the above apply to the fullest extent permitted. The parties agree these allocations of risk are reflected in the pricing and are an essential basis of the bargain.

11. Indemnification

You will defend and indemnify us against third-party claims, and resulting damages, costs, and reasonable attorneys’ fees, arising from your Customer Content, your applications and their actions, your use of the Service in breach of these Terms, or your violation of law or third-party rights.

12. Suspension and termination

You may stop using the Service at any time. We may suspend or terminate access with reasonable notice for convenience, or immediately for breach of these Terms, non-payment, security risk, or legal necessity. Upon termination we will, on request made within thirty (30) days, delete or return Customer Content in our possession, subject to legally required retention. Sections that by their nature should survive (including 3, 6, 7, and 9–14) survive termination.

13. Changes to the Service and these Terms

We may update the Service and these Terms as the product evolves. For material changes to the Terms we will give notice – for example by email or a notice in the dashboard – before they take effect. Continued use of the Service after the effective date constitutes acceptance. The current version is always available on this page.

14. Governing law and general

These Terms are governed by the laws of the State of Delaware, USA, excluding its conflict-of-law rules, and disputes will be resolved in the state or federal courts located in Delaware, whose jurisdiction both parties accept. If any provision is unenforceable, the remainder stays in effect. These Terms, together with the Privacy Policy and any order we agree with you, are the entire agreement regarding the Service. You may not assign them without our consent; we may assign them in connection with a merger, acquisition, or sale of assets.

Questions about these Terms: [email protected].

See also our Privacy Policy.