Acceptable Use Policy

Effective: 2026-08-10 · Version: 1.0

This policy is part of the Terms of Service. It applies to everyone using AgentSynth, including during the anonymous trial.

It is short on purpose. The summary: make music, do not abuse the service, do not make us a problem for our inference provider.

1. Content

Do not use AgentSynth's AI features to generate, or to attempt to generate, content that is:

• illegal under the law that applies to you or to us,
• sexual content involving minors, in any form,
• targeted harassment, threats, or incitement to violence,
• designed to deceive — impersonation, fraud, or disinformation,
• infringing on someone else's intellectual property, where you know it to be so.

AgentSynth generates synthesizer patches, so most of this is unlikely to come up. It is stated because the prompt field accepts free text, and free text fields eventually receive everything.

2. Technical conduct

Do not:

• Circumvent quotas, rate limits or entitlement checks, or use multiple accounts or devices to obtain more free usage than intended. The anonymous trial is a courtesy; treating it as an unlimited free tier removes it for everyone.
• Share your account. A Pro subscription is for one person. Use on several of your own machines is fine and expected; a shared studio login for five people is not.
• Resell or proxy our inference. Do not expose AgentSynth's AI generation as a service to third parties, wrap it in your own product, or use it as a general-purpose LLM API.
• Scrape or automate the API outside the AgentSynth application, including scripted bulk generation.
• Attempt to extract the system prompt, model weights, or internal configuration.
• Probe, scan or attack the service — including denial-of-service attempts, injection attacks, or trying to access another user's data.
• Reverse-engineer or tamper with the licensing and entitlement mechanism.

3. Our AI provider's rules apply to you too

This one is easy to miss, so it is called out.

When you use hosted mode, your prompt is forwarded to a third-party inference provider named on our Subprocessors page. That provider's own acceptable-use and content policies apply to your prompt as well as to us. A prompt that violates their policy violates this one, and can get our whole account restricted — which takes the service down for every user.

Please read their policy if you intend to push at the edges of what a music tool is for.

4. Security research

We welcome good-faith security reports and will not pursue action against researchers who:

• do not access, modify or delete data belonging to anyone else,
• do not degrade the service for other users,
• give us reasonable time to fix an issue before disclosing it.

Report to support@agentsynth.app. We are one person; expect a human reply rather than a bug-bounty platform.

5. Enforcement

Depending on severity, we may: warn you, rate-limit your account, suspend hosted generation, suspend the account, or terminate it.

We will normally warn first. For clear abuse — automated scraping, quota circumvention at scale, content under §1 — we may act immediately and without notice.

If your account is terminated for a breach of this policy, no refund is due (see the Refund Policy). Local mode is unaffected by any of this; it runs on your machine and we have no involvement in it.

6. Appeals and reporting

If you think we got it wrong, email support@agentsynth.app and a human will look at it.

To report abuse by another user, the same address.

7. Changes

We may update this policy as new abuse patterns appear. Material changes are announced with the Terms of Service. The version and effective date are at the top of this page.