Acceptable Use Policy
Effective 14 August 2026 · Version 3.0
NovaBuild gives every user real compute: agents that write code, containers that run it, and hosting that serves it. This policy sets the boundaries that keep that capability safe, lawful and available to everyone. It applies to prompts, generated projects, sandbox activity, deployed sites and custom domains.
Plain language summary
- Do not build or host content that is illegal, deceptive, exploitative or dangerous.
- Do not use sandboxes for mining, proxying, scraping at scale, load testing systems you do not own, or anything designed to consume compute for its own sake.
- Do not attempt to break tenant isolation or bypass metering. Report weaknesses instead.
- We enforce proportionately, from a warning to termination, and severe cases act immediately.
The summary is for orientation only. The numbered sections below are the operative text.
1. Purpose and scope
This policy is part of the Terms of Service. It applies to you, to every member of your workspace, to agents acting on your instruction, and to anyone using an application you deploy through NovaBuild. If you resell or embed NovaBuild output, you are responsible for the behaviour of your own users under this policy.
2. Core principles
- Be lawful in your own jurisdiction and in the jurisdictions you serve.
- Do not harm people, whether through content, code or automated conduct.
- Do not take more than your share of shared compute.
- Do not attack, probe or degrade the platform or other tenants.
- Be honest about what your application does and who runs it.
3. Prohibited content
You may not generate, store, host, link to or deploy content that:
- sexually exploits or endangers minors in any way, which we report to the relevant authority without notice;
- is non consensual intimate imagery, or sexual content produced from a real person's likeness without consent;
- incites or celebrates violence, terrorism, genocide or self harm, or provides operational instructions for weapons, explosives or biological, chemical, radiological or nuclear harm;
- harasses, stalks, doxxes, threatens, or targets a person or group with hateful abuse based on protected characteristics;
- is fraudulent or deceptive, including phishing pages, fake stores, fake support desks, romance scams, investment scams and impersonation of a brand, institution or person;
- distributes malware, ransomware, spyware, stalkerware, keyloggers, credential stealers or exploit kits;
- facilitates the sale of illegal drugs, stolen data, forged documents, hacked accounts or trafficked goods and services;
- constitutes unlicensed regulated activity, for example offering securities, gambling, lending, insurance, pharmaceuticals or weapons without the required authorisation;
- infringes copyright, trademark, patent, trade secret, database or publicity rights;
- presents AI generated media of a real person as authentic without clear disclosure.
4. Prohibited conduct
- Creating multiple accounts or workspaces to multiply free credits, evade quotas, or evade a suspension.
- Sharing a single account across a team instead of adding members, or reselling access to your workspace.
- Manipulating referral, gallery, template or credit systems, including self referral loops and fake signups.
- Reverse engineering platform prompts, extracting our internal system prompts for republication, or using the service to build a competing model of our orchestration layer from scraped internals.
- Misrepresenting NovaBuild as endorsing, reviewing, auditing or certifying your application.
- Using the service to make automated decisions about people, such as credit, employment, housing, insurance or benefits, without human review and lawful basis.
5. Security and platform integrity
Sandboxes are isolated for a reason. You may not attempt to escape a container, access another tenant's data, read orchestration metadata, tamper with billing or audit records, abuse service accounts, or interfere with rate limits and quotas.
Testing your own project is allowed
You may test, fuzz and attack applications you own inside your own sandbox. You may not use NovaBuild to attack anything you do not own or have written permission to test, and you may not attack NovaBuild itself outside the Responsible Disclosure policy.
6. Compute fair use and quotas
Fair use is enforced technically as well as contractually. Current controls include a cap on concurrent sandboxes per workspace, request rate limits on chat, build and deploy endpoints, per command wall clock limits, disk and memory ceilings inside the container, and an idle reaper that stops machines with no activity.
| Prohibited compute use | Why |
|---|---|
| Cryptocurrency mining or proof of work of any kind | Consumes shared capacity with no product purpose |
| Running a VPN, TOR relay, open proxy or traffic tunnel | Turns our egress into an anonymity service |
| Torrenting or hosting a file distribution endpoint | Bandwidth abuse and infringement risk |
| Load testing, stress testing or denial of service against any third party | Attack traffic from our address space |
| Bulk scraping of sites or APIs against their terms | Legal exposure and provider blocks |
| Long running background workers unrelated to building your app | Sandboxes are for build, preview and test |
| Model output farming at volume to resell tokens | Metering evasion |
If your legitimate workload needs higher quotas, email support@novabuild.dev and describe it. We would rather raise a limit for a real use case than have you work around it.
7. Model and agent misuse
- Do not use prompts designed to extract another user's data, our credentials, or provider system prompts.
- Do not attempt to disable guardrails, safety filters or audit logging in the platform.
- Do not use the agents to generate content in the prohibited categories in section 3, including code whose only purpose is abuse.
- Do not feed untrusted third party content into an agent with full autonomy and then blame the outcome on us. Treat external content as data, not instruction.
- Respect the model providers' own usage policies. A provider ban can force us to stop serving your workspace on that model.
8. Automation, scraping and bulk sending
You may automate your own workflow through documented endpoints within your rate limits. You may not send unsolicited bulk email, SMS or messages from a NovaBuild deployment, harvest contact details, or run outreach that breaches anti spam law such as CAN SPAM, CASL, the UK PECR or the Australian Spam Act. Transactional email from your app must include a real sender identity and an unsubscribe path where the law requires one.
9. Regulated and high risk uses
You may build in regulated spaces, and you carry the compliance burden. NovaBuild is not a HIPAA business associate, is not PCI certified for cardholder data storage, and is not designed for classified or export controlled material.
- Do not store protected health information, full payment card numbers, or government issued identity numbers in project files or prompts.
- Do not use generated output as the sole decision maker in medical, legal, financial, safety or employment decisions.
- If your application serves children, you are responsible for age assurance and parental consent.
- Tell us before running anything where downtime creates physical risk. Sandboxes and previews are not built for that.
10. Intellectual property
You must have the rights to everything you upload, and you must comply with the licences of open source packages your project installs, including attribution and copyleft duties. Notices of claimed infringement are handled under the Copyright and Trademark policy, which includes our takedown, counter notice and repeat infringer process.
11. Deployed applications and hosted content
When you publish, you are the publisher. Your deployment must not host prohibited content, must not mislead visitors about who operates it, and must carry its own privacy and cookie notices where required. Custom domains must be domains you control, and we may stop serving a domain whose verification lapses or which is used for abuse.
12. Reporting abuse
Report abusive projects, deployments or domains to abuse@novabuild.dev with the URL, a description, and any evidence. Report security weaknesses to security@novabuild.dev under the Responsible Disclosure policy. Report suspected exploitation of a minor to abuse@novabuild.dev with the subject line marked urgent, and we escalate immediately.
13. How we investigate
- We triage the report and assign a severity.
- We review the minimum data needed, which may include project metadata, deployment content and audit logs. We do not read private chat history unless it is necessary to assess severe abuse or we are legally required to.
- We may contact the workspace owner for an explanation, unless notice would increase harm or destroy evidence.
- We record the decision and the reason in the platform audit log.
14. Enforcement ladder
| Step | When we use it | Effect |
|---|---|---|
| Warning | First, low severity, likely accidental | Notice with a fix window |
| Throttle | Compute overuse or noisy automation | Reduced quotas until behaviour normalises |
| Content or project disable | A specific project, deployment or domain is offending | That item stops serving, rest of the workspace continues |
| Run suspension | Repeat breach or unresolved warning | Agent runs and deploys blocked, data stays readable and exportable |
| Workspace suspension | Serious or persistent breach | Access blocked pending review |
| Termination | Severe abuse, illegality, or repeated evasion | Account closed, no refund for the affected period |
Content that endangers people, clear illegality, active attacks and platform threats bypass the ladder and are actioned immediately.
15. Appeals and reinstatement
Every enforcement decision can be appealed by emailing appeals@novabuild.dev within 30 days with the workspace name and your explanation. A person who was not involved in the original decision reviews it, and we aim to answer within 10 business days. If we got it wrong we restore access and any credits consumed by the mistake. Reinstatement after a genuine breach requires the underlying content or conduct to be fixed first.
16. Changes to this policy
We update this policy as new abuse patterns appear. The effective date and version at the top reflect the current text, and material changes are announced in the app.
17. Contact
Abuse reports: abuse@novabuild.dev. Security: security@novabuild.dev. Appeals: appeals@novabuild.dev. Everything else: legal@novabuild.dev.
This document is published by the NovaBuild team as general information about the service. It is not legal advice, and it is not a certification or an independent audit of NovaBuild or of any third party. If anything here is unclear, or you need a countersigned copy for procurement, email legal@novabuild.dev.