Skip to content
Dagon Inference

Dagon Inference / Acceptable use policy

Acceptable use policy

Published · Last updated

Dagon Inference serves models with refusal behavior removed, so the responsibility for how they are used is yours. Use the service lawfully, and only against systems and people you own or have written permission to test. We suspend keys that do otherwise.

Welcome uses

  • Security research, reverse engineering and malware analysis on samples and systems you are entitled to examine.
  • Smart contract and protocol review of your own code and of code under an audit or bounty scope.
  • Red team work under a written agreement, capture the flag events and training.
  • General work with models that do not refuse legitimate questions.

Not allowed

  • Attacking, probing or disrupting systems, wallets, contracts or people without authorization.
  • Any sexual content involving minors, including fictional, text only and AI generated material, and any content that sexualizes minors.
  • Non consensual intimate imagery, and sexual content depicting real people without their consent.
  • Human trafficking; terrorism and violent extremism, including recruitment, planning and incitement; threats and doxxing; content that facilitates self harm.
  • Fraud, impersonation, harassment, or content to deceive victims, including scam and phishing campaigns against real targets.
  • Chemical, biological, radiological or nuclear weapons, and instructions to manufacture explosives or firearms or to modify them illegally.
  • Developing, deploying or operating malware, ransomware, botnets, credential stealers, phishing kits, wallet drainers or exploit code against systems, wallets or contracts you do not own or lack written authorization to test; automated attacks; credential stuffing; and using the service to exploit a live protocol for funds outside an authorized bounty or audit scope.
  • Laundering, sanctions evasion, scams and rug pulls, market manipulation, and deceptive investment or impersonation content.
  • Breaking sanctions or export laws that apply to you, or using the service from places where it is unlawful.
  • Abusing the service: evading rate limits, reselling or sharing access to evade limits, reselling keys to hide misuse, creating keys automatically beyond the rate limits, training competing models on outputs, or overloading the gateway.
  • Distributing malware.
  • Infringing intellectual property or other rights.

Responsibility and enforcement

Responsibility for what you send and how you use the outputs is yours. We do not monitor or review content. We may act on reports and other information we receive.

We may suspend or end a key, refuse requests, withhold or forfeit balances (see the credits section of the terms), preserve information and report to authorities. We may do so without notice where we think it necessary to prevent harm, to comply with law, or to protect our upstream relationship. We report apparent child sexual exploitation material to the NCMEC CyberTipline and to law enforcement where required.

We respond to valid legal process and emergency requests; see the privacy policy. We do not store prompts, so our ability to investigate is limited; do not read that as permission. Report misuse to abuse@inkwell.finance; see report abuse and legal notices for what to include.