Last updated: March 18, 2026
Ella Software, LLC ("Ella," "we," "us," or "our") is committed to using AI responsibly — especially in the context of education and child support. This page explains how AI works within Ella, what happens to your data, and the principles that guide how we build and deploy AI features.
Questions? Email us at hello@ella.kids.
Ella uses AI to help professionals and caregivers create visual supports for neurodivergent children. Users include special education teachers, therapists, school staff, and parents and guardians. When you describe an avatar in words, or generate a social story, schedule, worksheet, or other support, AI processes your input and produces the output you see.
Ella is designed so that AI models receive only what is needed to generate your content — nothing more.
No personally identifiable information is transmitted to AI models. Before any request reaches an AI model, we anonymize the input. The model never receives your name, email address, or any identifying information about you or any child.
We do not collect educational records, and we do not transmit them. Ella is not built on student data. The platform generates visual supports from the context a user provides — a situation, a routine, a behavior to model — not from academic records, IEPs, or any other educational data. There is nothing of that kind in our systems to transmit.
AI models do not retain your data. Data processed by our AI sub-processors is not retained beyond the immediate request. It is not stored, logged, or used for any other purpose.
Ella does not use your inputs or your generated content to train, fine-tune, or improve any AI model. This applies to everything you create on the platform: avatar descriptions, story prompts, schedule content, and any other material. Our AI sub-processors are contractually bound to the same standard.
AI-generated content on Ella is a starting point, not a finished product. Users are expected to review, edit, and apply their own judgment before using any visual support with a child. Ella does not make clinical, therapeutic, or educational decisions — all output is a tool to support the work of the professional or caregiver using it.
Safety. Ella's AI features include content guardrails designed to prevent generation of inappropriate or harmful material. All AI-generated content is filtered before it is presented to users.
Data minimization. We collect and share only what is necessary. The platform is designed so that effective visual supports can be created without student records, diagnoses, or sensitive child information — because we don't ask for any of it.
Fairness. AI models can reflect biases in their training data. We actively review Ella's outputs for quality and appropriateness, and we encourage users to flag content that seems off.
Transparency. We aim to be clear about what AI does and does not do in our platform. If our models, data practices, or capabilities change in a meaningful way, we will update this page and notify users.
Continuous improvement. We monitor AI performance, review user feedback, and update our approach as the technology and our understanding of it evolve. AI in educational and therapeutic contexts is an active area of development, and we take seriously our responsibility to stay current and respond to problems when they arise.
Ella uses third-party AI providers to power content generation features. All AI sub-processors are contractually bound to receive anonymized inputs only, not retain data beyond the immediate request, and not use any data for model training. A current list is available upon request at hello@ella.kids.
We may update this policy as our AI capabilities or provider terms evolve. The "Last updated" date at the top reflects the most recent revision. Material changes will be communicated via email or a notice on our website.
Ella Software, LLC hello@ella.kids
