Intellectual Property and AI-Output Licence

Last updated: September 30, 2026

1. Creator ownership

Stories, worlds, characters, and other original content on Mystrix are created by independent creators, and each creator retains ownership of their original content (Terms of Service §5). Publishing on Mystrix grants the operator the hosting, display, distribution, and promotion licence described in the Creator Agreement; it does not transfer ownership of the underlying work.

A creator may upload images or explicitly request generation or refinement with an OpenAI, Google Gemini, or Z.ai key entered for the open authoring session, or use the Mystrix image service where enabled. The creator must have rights to the submitted brief and references; generated output may still resemble third-party material and must be reviewed before use. Private candidates and retained composition sources are not published automatically. Mystrix may also apply approved reviewed house-owned images to eligible Story surfaces. Creator uploads and private generated sources remain under creator custody and are never made cross-Story stock. An approved reviewed house-owned discovery or surface image may be reused only when its visual specification, role, rights evidence, and safety review match. Each applied full-frame scene image remains restricted to the exact owning account, Story version, and event.

2. AI choreography and session presentation

When you play a Story, the platform uses AI to select and choreograph fixed creator-authored material in response to your input. AI does not write live Story prose. The resulting session presentation is provided for your personal use within the platform (Terms of Service §5). This licence does not extend to redistributing, republishing, or commercially reusing the session presentation or the creator's underlying Story content outside Mystrix without permission from the rights holder.

3. Your own input

The text you write into a session, a world chat, or a creator draft remains subject to this licence and to the Privacy Policy's description of how it is processed (see §4 below); using the platform does not require you to give up more rights in your own writing than are necessary to let Mystrix operate the Service you asked for — choreographing an authored Story turn, storing the session so you can resume it, and the other uses described in the Privacy Policy.

4. Third-party AI processing

Story choreography is primarily processed by DeepSeek using deepseek-flash, with Groq as its configured distinct fallback. These services select and arrange fixed creator-authored material; they do not write live Story prose. General platform text, including support questions and non-story world chat, remains produced on Mystrix's own infrastructure through loopback-only Ollama. OpenAI, OpenRouter, and Alibaba Cloud (Qwen) remain approved hosted providers but are not part of the canonical reader route. If a creator supplies their own text-generation credential, authoring requests using it remain pinned to that creator-selected provider and do not fall through to the platform reader route. Mystrix does not control whether a third-party provider retains or trains on text it receives; that is governed by that provider's own terms.

DeepSeek is the deepseek-flash text provider and also performs the separate, limited safety-verification step below; it is never used to generate or edit an image. Explicit Creator image requests instead use a session key for OpenAI, Google Gemini, or Z.ai, or the Mystrix image service where enabled.

• Safety review of uploaded visual media. During classifier-backed processing of images and videos you upload, a reduced image derivative or, for video, one extracted poster frame is sent to DeepSeek for automated safety verification. DeepSeek uses those pixels only to return safety scores; it does not generate or edit images. OpenAI does not receive those derivatives or poster frames. If a derivative cannot be prepared, DeepSeek is unavailable, or the result is flagged, the uploaded asset remains unavailable for Story application until human review records an explicit approval; rejection keeps it blocked. Only approved reviewed Creator uploads or approved reviewed house-owned assets can be applied. Automated verification is a moderation aid, not a guarantee that every problematic item is caught.

Story speech is generated by Mystrix itself: the sanitized spoken prose, its language, and the selected in-house designed voice identity are processed on Mystrix's own voice box, and Mystrix stores each admitted result under the exact speech recipe and composes continuous Story playback from those durable clips. Repeated playback and matching recipes use stored audio rather than synthesizing the line again. No hosted speech provider receives your Story prose. If a render's outcome is unknown, the next preparation attempt uses the same content key, so a finished render is collected rather than produced again. Optional creator sound-effect generation also runs on the voice box: only the effect description you write (at most 500 characters) and a requested duration of 10 seconds or less are processed there, never by a third party; Mystrix measures the generated audio and stores it as your ordinary sound asset, and you decide whether to share that asset with other creators through the explicit reuse grant.

Mystrix does not control whether those third parties retain or train on what they receive; that is governed by their own terms.

5. Infringement claims

If you believe content on Mystrix — including AI-generated output — infringes your intellectual property rights, contact us through our support system and identify the content and the right you believe is infringed. See Moderation, Takedown, and Appeals for how a takedown request or content report is handled today; this document does not claim a separate, formally designated infringement process beyond that.

6. Contact

Questions about content rights or this licence should go through our support system.