Terms of Service

These Terms of Service (the “Terms”) are a contract between you and Novix (“Novix,” “we,” “us”). They govern your access to and use of the Novix application, dashboard, APIs, command line tools, connected doorways, and related services (the “Service”). By creating an account, connecting an integration, or using the Service, you agree to these Terms.

If you accept these Terms on behalf of a company or team, you represent that you are authorized to do so, and “you” means that organization and its authorized users. You must be at least 18 and able to form a binding contract to use the Service. The Service is built for businesses and is not directed to consumers or to children.

Novix is actively developed. Features change and availability is not guaranteed, so the sections on warranties and liability below apply in full.

Novix reads support tickets from the tools you connect (help desks, chat, email, issue trackers, and error and observability services), classifies and diagnoses them, and, when it concludes a ticket is a genuine code defect, drafts a candidate fix. It can also draft a suggested reply to the customer and hand a diagnosis to your engineers.

Novix is a tool that assists your team. It is not a law firm, not a security auditor, and not a substitute for your own engineering judgment. It does not take independent action on your production systems beyond the specific, scoped actions you connect and authorize (for example, opening a pull request on a repository you connect).

This is the most important thing to understand about the Service. The fixes Novix produces are drafts and suggestions, generated by an AI model. Novix builds and tests them before they reach you, and that is still not a guarantee that they are correct.

• Built and tested, never on your systems. Novix clones your connected repository into an isolated, single-use container at our provider, applies the drafted patch there, and runs your repository’s own build and test commands against it. This happens on every drafted fix. It never runs on your infrastructure and nothing is written back to your repository from that container. See the privacy policy for what that container receives and how long it lives.
• A passing build is not a correctness guarantee. It means your build succeeded and the tests you have written passed. It says nothing about the tests you have not written, and a fix can compile, pass, and still be wrong or insecure. Treat every diff as a proposal that a human must read, whatever the result says.
• The result never decides anything. A failed build is recorded on the ticket for you to see; it does not stop a pull request being opened and does not stop you approving one. Novix may also be unable to build a given patch at all (for example, if it cannot detect a build command in your repository), in which case it says so on the ticket rather than implying it checked.
• Written from your real files, or not at all. Novix drafts a diff only when it could read the relevant files out of your connected repository. When it cannot read them, it drafts nothing and says so on the ticket, rather than guessing at a patch against files it has never seen.
• No warranty of correctness. AI models can be wrong, incomplete, or confidently mistaken. A diagnosis may misidentify the root cause; a suggested fix may be incorrect, insecure, or introduce new problems. You must not assume any output is correct.

Novix surfaces a confidence signal and a risk rating with its suggestions to help you triage. These are aids to your judgment, not guarantees, and they do not shift responsibility for the change onto Novix.

By default, a drafted fix is opened as a draft pull request that waits for a person to review it. Nothing is merged to your codebase or deployed to your systems by Novix. You retain full control of your version-control workflow, your reviews, your CI checks, and your deploys.

A drafted fix always opens as a pull request. The only thing Novix decides on its own is whether it arrives as a draft or as ready-for-review, and it arrives ready-for-review only when its internal fix review ran and passed at low risk, on a genuine defect rather than a requested change, with nobody having already decided the fix and the workspace inside its spend cap. If any of that fails it opens as a draft carrying the reason. Either way it never merges code and never deploys. A human still approves and merges. Direct-to-production merging is not a feature of the Service.

An approval can be given from more than one place: the dashboard, a button in Slack, a comment on a connected tool, a reply to an email, or by merging the pull request itself. They are the same act with the same consequences, including billing. See plans, fees and billing.

Because you review and approve every change, you are responsible for what you merge. You agree not to rely on Novix as the sole check on a change to your software.

When you use Novix, you agree to:

• Review, test, and independently validate any suggested fix before merging or deploying it, using your own code review, testing, and release processes.
• Maintain your own backups, version control, and the ability to roll back changes. You should be able to revert any merge.
• Only connect code, tickets, and systems that you have the right to connect, and only grant Novix the access scopes you intend. We recommend least-privilege, read-only access wherever it is enough.
• Set and monitor your own spending limits with every third party you connect, including your inference provider. See bring your own inference key.
• Keep your account credentials and API keys secure, manage who on your team has access, and promptly remove access for people who leave.
• Tell us promptly if you believe your account, a key, or a connected credential has been compromised, or if you notice activity on your workspace you did not authorize.
• Comply with all laws and with the terms of the third-party services you connect through Novix.

The Service works by connecting to third-party tools you choose (for example, Zendesk, Intercom, Slack, Sentry, Datadog, GitHub, and others). You are responsible for your use of those services and for the credentials, tokens, and access scopes you provide to Novix.

Your relationship with each third-party provider is governed by that provider’s own terms. Novix is not responsible for third-party services, their availability, their pricing, or their handling of your data. You can disconnect any integration at any time; doing so deletes the stored credentials for that integration.

Connecting a repository authorises Novix to build it. When you connect a code repository, you instruct us to clone it into an isolated container and run that repository’s own build and test commands there, in order to check a drafted fix before you approve it. Those commands are your code and they execute with the network access the container provides, so only connect a repository you are willing to have run this way, and make sure you have the right to authorise it. You can stop this at any time by disconnecting the repository. It is a step of the analysis and there is no separate setting that turns it off while a repository stays connected.

Disputes with a third party are not disputes with us. To the fullest extent the law allows, you release us from claims arising out of a dispute between you and a service you connected, and out of what somebody else with access to your workspace did in it. If you are a California resident, you give up California Civil Code section 1542, which would otherwise keep a release from covering claims you did not know about when you gave it.

Novix runs on large language models. By default those calls run on our inference account and we pay for the tokens. A workspace on a plan that allows it can point Novix at its own inference account instead, and from that moment every model call for that workspace runs on your account and bills to you. Today you can point it at Anthropic, OpenAI, Google Gemini, DeepSeek, OpenRouter, or Kimi. Anything else is an endpoint we have agreed with you in writing.

What you are taking on

• All of it, however it happened. That includes tokens spent on runs you started, on runs your connected tools started for you, on tickets your own customers filed, on retries, on failed or abandoned runs, on work Novix did that you never used, on duplicate or automated ticket volume, and on spend caused by your own configuration, by a third-party service you connected, or by a defect in the Service.
• We cannot cap your provider spend. Nothing in Novix limits what your provider will bill you. Set your own budget, alerts, and hard limits in your provider’s console before you paste a key. Novix’s own analysis limits exist to protect the Service, they can change at any time, and you must never rely on them as a spending control.
• Any cost figure we show is an estimate. Usage numbers in the dashboard come from our own metering, are counted differently from your provider’s, can be wrong, and are never a bill, a quote, or a commitment. Your provider’s invoice is the only authority on what you owe them.
• The provider relationship is yours. Their terms, their acceptable use policy, their rate limits, their outages, their price changes, their model deprecations, and their handling of your data are between you and them. We are not a party to it, not a reseller, and not your agent. See the privacy policy for what it means for your data.
• Some providers cost more per run. The engine is built and tuned for Anthropic’s models. On providers that do not support prompt caching, Novix re-sends its working context every turn instead of reading it back at a discount, so the same work can cost you materially more, and some cost and quality controls the engine uses are not available at all. We state this in the app at the point you choose. It is how those APIs work, not a defect in the Service.

What we do with the key

• It is encrypted at rest under a key held outside the database, and only its last four characters are ever shown back to you.
• It is sent only to the provider endpoint bound to it, to run your own workspace’s model calls. We send one cheap test request when you save it, so a key that does not work is refused at the door rather than three stages into a diagnosis.
• By saving a key you confirm you are authorised to use it, authorised to route your ticket content and code context to that provider, and that doing so breaches no agreement of yours.
• Rotate or remove it whenever you like. Removing it takes effect on the next model call, and from that point your workspace runs on our account and approved fixes are metered as described in plans, fees and billing.

If your key stops working

When your provider refuses your key, Novix stops. It deliberately does not fall back to our account, because that would move your token bill onto us at the exact moment your configuration broke. Runs on that workspace fail until the key is fixed, and we email your workspace administrators to tell them. That behaviour is intended and is not a failure of the Service.

If you point us at an endpoint you or a third party operate, you are responsible for that endpoint, for what it does with your key and your data, and for anything it returns. We screen outbound requests against non-public network addresses, and that screen protects our infrastructure. It is not an assurance about a host you chose.

The Service enforces limits: how many analyses a workspace may run in a day, how much a brand new workspace may run in its first day, how fast an endpoint may be called, how many accounts may be created from one place, and how much metered spend a workspace may reach. Those numbers are set by us to keep the Service running and the bill survivable. We can change them at any time and without notice. They are not a commitment about capacity, and they are never a budget control on your behalf.

Use the Service for your own bugs, in your own software, at a human scale. A workload that consumes resources far out of proportion to what you pay, whether or not it trips a specific limit, is something we may throttle, cap, or price separately after telling you.

Costs you cause, you pay. If you use the Service in breach of these Terms and that causes us to incur inference, sandbox, storage, egress, or other infrastructure costs, you agree to reimburse those costs at our then-current rates, on top of anything else we may do about it.

A bug in your favour is not a licence. If a defect, a misconfiguration, a promotional error, or a missing check gives you access, capacity, seats, or pricing you did not pay for, you may not use it or keep using it. Tell us. We may correct it, including retroactively, and invoice the difference.

You agree not to do any of the following, and not to let anyone else:

Use it on things that are not yours

• Build, diagnose, or modify software you do not have the legal right to access or change.
• Upload, connect, or process data you are not permitted to share, or use the Service in a way that violates the rights of others.
• Use the Service for anything unlawful, or to create or distribute malware, exploits, or code intended to cause harm.
• Share login credentials or bypass workspace permissions.
• Resell, sublicense, rent, or operate the Service on behalf of a third party, use it to build or improve a competing product, or reverse engineer any part of it.

Get around a limit

• Evade, disable, or work around any rate limit, quota, plan limit, spend cap, trial length, payment requirement, or other technical limit.
• Create multiple accounts, workspaces, or organizations to get extra trials, extra free inference, extra seats, or to get back in after we removed you. Signing up with disposable or programmatically generated addresses for that purpose is included.
• Forge, spoof, or manipulate anything the Service uses to identify a caller, including source addresses, forwarded headers, edge headers, device signals, session cookies, or webhook signatures.
• Automate account creation, or point scripts, headless browsers, or bots at surfaces meant for people.
• Generate tickets, tasks, or analyses in bulk or at machine speed in order to consume our inference, our sandbox capacity, or anyone else’s allowance.

Attack the Service

• Probe, scan, or test the vulnerability of the Service or the systems around it, except as allowed in the next section.
• Interfere with or degrade the Service, including denial of service, flooding, or load testing without our written permission.
• Attempt to reach an account, workspace, ticket, repository, key, or any data that is not yours.
• Exploit a bug rather than report it, or use a credential, token, or session you obtained by any route other than us giving it to you.

Abuse the build container

Novix clones your connected repository into an isolated container and runs that repository’s own build and test commands there. That container exists to check a drafted fix and nothing else. You must not use it, or cause it to be used, as general purpose compute, to mine cryptocurrency, to host or proxy network traffic, to scan or attack any third party, to attempt to persist beyond the run, to attempt to reach our infrastructure or another customer’s data, or to move data out of it. You are responsible for what the scripts in your repository do when they run.

Attack the model layer

You must not use content Novix reads, whether ticket text, attachments, filenames, comments, repository files, or anything else, to try to make the Service act outside your workspace: to extract our prompts or model configuration, to reach another customer’s data, to make Novix take an action you are not entitled to take, or to get around the human approval boundary.

To be clear about what this does not ban. Our customers build software, some of it built on AI, and “a user can type instructions into our product and our prompt breaks” is a real bug report. Filing that ticket, with the payload in it as evidence, is exactly what Novix is for and is always allowed. The line is intent and target: an attack aimed at our systems is a breach of these Terms, a report about your own product is a ticket.

We may investigate, and we may throttle, suspend, or terminate access for a violation of this section, with or without notice depending on the risk. See suspension.

We would rather hear about a hole than read about one. If you think you have found a security vulnerability in Novix, email support@getnovix.ai with enough detail to reproduce it. We will not pursue a report made in good faith under the conditions below. We do not run a paid bounty programme today.

• Test only against your own workspace and your own data, never another customer’s.
• Stop at proof. Do not read, copy, change, or keep data that is not yours, do not degrade the Service, and do not run automated scanning, load testing, or brute force.
• Do not use a finding to gain paid capacity, extend a trial, or avoid a charge.
• Give us a reasonable chance to fix it before you tell anyone else.

Anything outside those conditions is not authorised testing, and the acceptable use section applies to it in full.

Novix is organized into workspaces (teams). An account may belong to a workspace with roles such as administrator and member. Administrators can connect integrations, change auto-summon settings, invite and remove members, and change roles. You are responsible for the actions taken under your account and by members of your workspace, including approvals, merges, and anything they connect.

You must provide accurate account information and keep it current. You are responsible for maintaining the confidentiality of your login credentials and for all activity under your account. One login belongs to one person. Tell us promptly if you believe your account has been compromised.

Free trial

New workspaces get a 7-day free trial of the full Service, starting when the workspace is created. No payment method is required to start one and nothing is charged during it. We may change the trial length for future signups; a change never shortens a trial already underway. When a trial ends without a paid plan, the workspace moves to our free plan and keeps running at that plan’s limits; your data is kept and nothing is charged. A trial never converts to a paid plan by itself; you subscribe when you choose to. See how the trial works for the plain-English version.

Workspace plans

We offer a free plan and a paid plan, each for one workspace with unlimited people, as shown on the pricing page. The paid plan is billed monthly in advance at a flat price per workspace. The free plan carries no subscription and no recurring charge; the only fees on it are the per-fix fees described below. Inviting someone or changing their permissions does not change the price of either plan.

A payment method is required before the first approved fix. Where a workspace runs on our inference account, we collect a card before anyone approves a fix. On a plan with no subscription, adding the card creates a billing arrangement with no recurring charge; it exists so the per-fix fees below can be collected, and nothing is charged at the moment you add it. Until a card is on file we refuse approvals that would bill.

Approved fixes

When a workspace runs on our inference account, each approved fix is billed at the per-fix price shown in the app, from the first one, with no limit on the number of fixes you may approve in a billing period. On a plan that carries no subscription we apply a hard cap on approved-fix fees in the same billing period, in an amount stated in the app; at that cap, further approvals are refused until the period rolls over. When a workspace runs on its own inference key, approved fixes are unlimited and nothing is metered, because you already paid your own provider for the tokens.

Whose key paid is decided per run, not per setting. Tokens are spent while Novix works a ticket, which is usually well before anyone approves the fix. We meter against the account that actually paid for that run, taken from our own usage records, rather than against whatever the settings page says at the moment of approval. Pasting a key after the work is done does not make that fix free, and clearing one does not charge you for work your own provider already billed you for.

Every approval meters, wherever it happens. Approving in the dashboard, pressing the button in Slack, commenting the command on a connected tool, replying to the email, and merging the pull request Novix opened are the same act. Merging is an approval. If you do not want a fix, close the pull request instead of merging it.

Payment, taxes and refunds

• Payments are processed by Stripe. Card details go to Stripe and never touch our systems.
• Fees are exclusive of taxes. You are responsible for any sales, use, VAT, GST, or withholding taxes, other than taxes on our income.
• Fees are non-refundable except where the law requires otherwise. Cancelling stops the next renewal; it does not refund the current period, and we do not pro-rate a partial month.
• A trial never turns into a charge on its own. No card is on file until you deliberately add one, and a trial that ends moves the workspace to the free plan rather than charging it. On a paid plan the first charge happens when you subscribe; cancel before a renewal date to avoid the next one. On a plan with no subscription, adding a card starts no recurring charge and the first charge is the first fix you approve.
• If a payment fails we may retry it and may suspend the workspace until it clears. Fees continue to accrue during a suspension for non-payment.
• We may change prices. A change takes effect at your next renewal after we tell you, and your remedy if you do not like it is to cancel before then.
• Query a charge in writing within 30 days of the invoice or it is treated as accepted. Raise it with us before you raise a chargeback; a chargeback on undisputed fees is a breach of these Terms and we may suspend the workspace and recover our costs.

Third-party costs are never ours. Your inference provider, your cloud, your CI minutes, your version-control host, your help desk, and every other service you connect bill you directly under your own agreements with them. Nothing you pay us covers any of it, and nothing in these Terms makes us responsible for it. See bring your own inference key.

We may suspend a workspace, an account, or a single feature immediately and without notice where we reasonably believe it is necessary: a breach of the acceptable use section, a security or abuse risk to the Service or to another customer, unusual spend or volume that looks automated, a legal requirement, or non-payment. Where the risk allows it we will tell you first, and we will tell you afterwards either way.

Suspension is not termination. We restore access once the cause is resolved. We do not credit fees for a suspension caused by your breach or by non-payment. If you think a suspension is wrong, email us and we will look at it.

Your content stays yours. As between you and Novix, you own your code, your tickets, and the data you connect (“Your Content”), and you own the code that results from a change you choose to merge. You grant Novix a limited license to process Your Content solely to provide the Service to you (for example, to diagnose a ticket and draft a fix).

We keep our product. Novix, including its software, models, prompts, and interfaces, and all related intellectual property, belongs to us. These Terms do not grant you any rights in the Service except the right to use it as described here.

What you may do with it. While these Terms are in force and your fees are paid, we grant you a limited, non-exclusive, non-transferable, revocable right to use the Service for your own business. That is a right to use the running product, not a licence to our code, and it ends when your access does. What you may not do with it is in acceptable use.

AI output. As between you and us, the diffs, diagnoses, and replies Novix drafts for you are yours to use, and we claim no ownership of them. We also make no warranty about them: an AI model can produce output that resembles someone else’s work, that carries a licence you did not mean to take on, or that is not protectable at all. Checking output for licence, security, and originality before you merge it is part of the review you already agreed to do.

Feedback. If you send us feedback or suggestions, we may use them to improve the Service without obligation to you.

Product analytics. We measure how workspaces get started with Novix and how the Service is being used, to run and improve it. These are counts calculated from records the Service already keeps, not a tracking script, and they exclude the content of your tickets. Our Privacy Policy describes exactly what is included, and you can email support@getnovix.ai to have your workspace excluded from it.

Working together means each of us learns things about the other that are not public. Your code, your tickets, your customers’ messages, your configuration and anything you connect are confidential, and so is anything either of us marks as confidential or that is obviously so, such as pricing we quote you outside the published plans.

We use your confidential information only to provide, secure and improve the Service, as our Privacy Policy describes. We do not disclose it outside Novix except to the subprocessors listed at subprocessors, each of which is under a written contract, and where the law compels us. If we are compelled, we tell you before we hand anything over unless we are legally forbidden to. We protect it with at least the care we use on our own.

None of this covers information that is already public without either of us breaking a duty, that you or we already had before it was shared, that comes from someone free to share it, or that either of us works out independently. These obligations run while you use the Service and for three years after, and for trade secrets they run for as long as the law protects them.

The Service is provided “as is” and “as available,” without warranties of any kind, whether express, implied, or statutory. To the fullest extent permitted by law, we disclaim all warranties, including implied warranties of merchantability, fitness for a particular purpose, title, and non-infringement.

Without limiting the above, we do not warrant that:

• any diagnosis, fix, reply, confidence signal, risk rating, or cost figure will be accurate, complete, secure, or fit for your purpose;
• the Service will be uninterrupted, timely, or error-free, or that defects will be corrected. We do not commit to an uptime level or a support response time unless we have signed one with you;
• a build or test run in our container proves anything beyond that your build succeeded and your existing tests passed, and no part of the Service is a security audit, a penetration test, or a compliance control;
• any particular model, provider, feature, or limit will keep working the way it does today. Models get deprecated, providers change their APIs, and we may change which model a stage runs on.

Features may change, be added, or be removed. You use the Service, and any output it produces, at your own risk and subject to your own review.

To the fullest extent permitted by law, Novix and its affiliates 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, arising out of or related to the Service or these Terms, even if we have been advised of the possibility of such damages.

This includes, without limitation, damages arising from a suggested fix, diagnosis, or other output that is incorrect, incomplete, insecure, or that you merged or deployed. Because you review and approve every change, decisions about your software are yours.

To the fullest extent permitted by law, our total aggregate liability arising out of or related to the Service or these Terms will not exceed the greater of the amount you paid us for the Service in the twelve months before the claim, or one hundred U.S. dollars (US$100). Amounts you pay a third party are not amounts you paid us and never count toward that figure.

These limits apply to every theory of liability, in contract, tort, negligence, strict liability, or otherwise, and apply even if a limited remedy fails of its essential purpose. They do not limit what you owe us in fees, your obligations under usage limits and fair use, or your indemnification obligations below. This allocation of risk is a fundamental part of the bargain between us and is reflected in the price. Some jurisdictions do not allow certain limitations, so parts of this section may not apply to you.

You agree to defend, indemnify, and hold harmless Novix and its affiliates, and their officers, employees, and agents, from and against any claims, damages, liabilities, fines, and expenses (including reasonable legal fees) arising out of or related to:

• your use of the Service and Your Content;
• the changes you choose to merge or deploy, and anything they cause;
• the code, scripts, and dependencies in a repository you connect, including what they do when they run in our build container;
• charges on any third-party account you connect, including your inference provider;
• a claim by one of your own customers or end users about data you processed through the Service;
• your violation of these Terms, of any law, or of the rights of a third party.

We will tell you about a claim promptly, let you control the defence with counsel of your choice, and cooperate at your expense. You may not settle a claim in a way that admits fault on our behalf, imposes any obligation on us, or fails to release us fully, without our written consent. We may take part in the defence with our own counsel at our own cost.

You may stop using the Service and close your account at any time. We may suspend or terminate your access if you violate these Terms, if we reasonably believe it is necessary to protect the Service or other users, or if we discontinue the Service. We may modify or discontinue the Service, in whole or in part, at any time.

On termination, your right to use the Service ends and any fees you owe become due. Export anything you want to keep before you close an account; we are under no obligation to hold your data after termination. Sections that by their nature should survive, including intellectual property, confidentiality, acceptable use, disclaimers, limitation of liability, indemnification, and disputes, will survive. Our handling of your data after termination is described in our Privacy Policy.

Talk to us first. If something goes wrong, email support@getnovix.ai with a short description of the problem and what you want. We will do the same if the complaint is ours. Neither of us may start an arbitration until 30 days after that notice, and that time is for trying to sort it out.

Binding arbitration. If it is not sorted out, any dispute arising out of or relating to these Terms or the Service will be resolved by final and binding arbitration, administered by the American Arbitration Association under its Commercial Arbitration Rules, before a single arbitrator, in English. This agreement to arbitrate is governed by the Federal Arbitration Act. Judgment on the award may be entered in any court with jurisdiction. Because these Terms do not yet name a governing law or a seat (see below), the arbitrator determines the locale under the AAA rules unless we have agreed one with you in writing.

What is carved out. Either of us may bring a qualifying claim in small claims court, and either of us may ask a court for an injunction or other equitable relief to stop unauthorised access, misuse of the Service, a breach of the acceptable use section, or infringement of intellectual property, without arbitrating first.

Who decides what. The arbitrator decides everything about a dispute, including whether these Terms cover it and whether this section applies to it. The one exception is the class waiver above: whether that paragraph is enforceable is for a court, not the arbitrator.

Costs. Each of us pays our own legal costs. How the arbitration’s own fees are split is decided by the AAA rules and by the arbitrator, who may award costs where the law allows it.

One year. Any claim must be brought within one year of when it arose, or it is waived, to the extent the law allows.

We may update these Terms as the Service evolves. When we make a material change, we will update the effective date above and, where appropriate, give notice. If you continue to use the Service after a change takes effect, you accept the updated Terms.

• Entire agreement. These Terms, the Privacy Policy, and any written agreement we sign with you are the whole agreement about the Service and replace anything said before. Where a signed agreement conflicts with these Terms, the signed one wins for the part that conflicts.
• Severability. If part of these Terms is unenforceable, the rest stays in force and that part is narrowed to what is enforceable.
• Waiver. Not enforcing something once does not waive it. A waiver counts only in writing.
• Assignment. You may not assign these Terms without our written consent, except to a successor of your business that is not a competitor of ours. We may assign them to an affiliate or in connection with a merger, financing, or sale of assets.
• Notices. We give notice by email to a workspace administrator or in the app, and either counts. You give notice to support@getnovix.ai. Notice is effective when sent.
• Electronic communications. Everything between us happens electronically: by email, in the app, and on these pages. You agree that counts as writing wherever the law asks for something in writing or signed, and that an agreement you accept by clicking is as binding as one on paper. Keep a copy of anything that matters to you; we do not keep one for you.
• Publicity. We will not use your name, your logo, or the fact that you are a customer in our marketing without your written permission. If you give it, you can withdraw it and we will take the mention down.
• Copyright and trademarks. Novix, the Novix mark, and everything on our sites are ours or our licensors’. Nothing here lets you use them, beyond saying factually that your team uses Novix. Other companies’ names and marks appear on our sites only to say what Novix connects to, and they belong to those companies.
• Accessibility. We build Novix towards WCAG 2.1 level AA. Every colour pairing in our design system is recomputed against the standard’s contrast formula on every build and the build fails if one falls below its bar. We have not had an independent accessibility audit, and we are not going to imply we have. If something in Novix is unusable for you, email support@getnovix.ai, tell us what and where, and we will fix it.
• Force majeure. Neither of us is liable for a delay or failure caused by something outside our reasonable control, including an outage at a provider we depend on, a network failure, a change to a third-party API, or a legal order. Payment obligations are not excused.
• Export and sanctions. You confirm you are not located in, or acting for, a country or person subject to United States sanctions or export restrictions, and that you will not use the Service in breach of them.
• Relationship. We are independent contractors. Nothing here creates a partnership, agency, employment, or fiduciary relationship, and neither of us may bind the other.
• No third-party beneficiaries. These Terms are between you and us. Nobody else gains rights under them.
• Jury trial. For any claim that does end up in court, both of us waive a trial by jury, to the extent the law allows.
• Headings. Headings are for reading, not for construction. “Including” means including without limitation, and these Terms are not read against whoever drafted them.

These Terms do not yet name a governing law or a court venue. If your team needs that settled before you can sign off, email us and we will agree it with you in writing. The disputes section applies in the meantime and rests on the Federal Arbitration Act, which does not depend on that choice.