AI & Data Safety
iMothership uses selected AI service providers to process requests initiated through our platform. We use business or API services whose published policies protect API inputs and outputs from general model training, subject to the account, plan and feature requirements below.
Provider terms can change. We review these controls regularly and configure conditional services to prevent training or unnecessary retention where the provider offers those controls. Last reviewed: 17 August 2026.
Google Workspace and Photos API Limited Use
The use of raw or derived user data received from Google Workspace or Google Photos APIs by iMothership will adhere to the Google API Services User Data Policy and the Limited Use requirements.
Our Limited Use commitments include:
- We access only the Google user data needed to provide user-requested, visible features.
- We do not sell Google user data or use it for advertising, credit decisions, or data-broker activities.
- We do not use or permit the use of raw, aggregated, anonymized, or derived Google Workspace or Photos data to create, train, or improve foundational or generalized AI/ML models.
- Google user data is transferred to an AI provider only to perform a user-requested feature and only through the provider's business or API service under a no-generalized-training configuration.
- Consumer AI endpoints and free service tiers that may use submitted content for generalized model training are not approved for processing Google Workspace or Photos user data.
- Where a provider offers an opt-out or Zero Data Retention control, that control must be enabled before Google user data may be processed by that provider.
Third-party service configurations
- Anthropic: commercial API.
- Google: billing-enabled Gemini or Vertex AI API.
- OpenAI: API platform business data controls.
- Replicate: authenticated API; downstream models are disclosed below.
- Fireworks AI: authenticated API with no-training data handling.
- Groq: GroqCloud API with data controls.
- ElevenLabs: API with model-improvement data sharing disabled.
- SpaceX AI / xAI: developer API using dedicated API keys and Zero Data Retention when Google user data is processed.
Multi-model services and downstream models
Replicate is used as a model-hosting platform for media-generation features. Current downstream models include Black Forest Labs FLUX Schnell, MiniMax Video-01, ByteDance PuLID, Clarity Upscaler, and Meta MusicGen. Some projects can also use a model selected in that project's configuration. The same Limited Use restrictions apply to every selected downstream model; a model may not receive Google user data if its terms or configuration permit generalized model training.
Self-hosted and offline models
When iMothership uses a self-hosted or offline model, execution takes place within private, isolated infrastructure. Google Workspace or Photos user data is processed locally and is not transmitted back to the model provider for training, model improvement, advertising, or any other secondary purpose.
Microsoft 365 and Outlook API Data Use
The use of raw or derived user data received through Microsoft Graph, Outlook, Exchange, or Microsoft 365 APIs by iMothership will comply with the Microsoft APIs Terms of Use, the Microsoft Privacy Statement, and applicable privacy and data-protection requirements.
Our Microsoft API data commitments include:
- We request only the minimum Microsoft API permissions and data needed to provide user-requested application features.
- We use Microsoft API data only within the permissions expressly granted by the customer or end user.
- We do not sell, redistribute, or use raw, aggregated, anonymized, or derived Microsoft API data for advertising, marketing, or data-broker activities.
- We do not use or permit the use of Microsoft 365, Outlook, Exchange, or Microsoft Graph data to create, train, or improve foundational or generalized AI/ML models.
- Microsoft API data is transferred to an AI provider only to perform a user-requested feature and only through the provider's business or API service under a no-generalized-training configuration.
- Where a provider offers an opt-out or Zero Data Retention control, that control must be enabled before Microsoft API data may be processed by that provider.
- We apply appropriate retention and deletion controls and delete Microsoft API data when it is no longer required for the requested feature or when access is withdrawn, subject to applicable legal obligations.
Self-hosted and offline processing
When iMothership uses a self-hosted or offline model, Microsoft 365 and Outlook data is processed within private, isolated infrastructure. It is not transmitted back to the model provider for training, model improvement, advertising, or any other secondary purpose. The third-party service configurations and downstream-model restrictions listed above apply equally to Microsoft API data.
Managing Microsoft access
Users and administrators can review or revoke iMothership's Microsoft access through Microsoft My Apps or, for personal Microsoft accounts, through Microsoft account consent management.
EU AI Act Transparency and Responsible Use
iMothership is committed to developing and using AI in accordance with applicable requirements of the EU Artificial Intelligence Act (Regulation (EU) 2024/1689), including its risk-based governance, transparency, human-oversight, and AI-literacy principles.
Our EU AI Act commitments include:
- We inform users when they are interacting directly with an AI system unless that interaction is obvious from the context.
- We identify or label AI-generated or manipulated content where required, including applicable deepfake and public-interest content disclosures.
- We assess AI features according to their intended purpose and risk, and do not permit prohibited manipulative, exploitative, discriminatory, or unlawful uses.
- We maintain meaningful human oversight for consequential uses and do not present AI output as a substitute for professional judgment where a person's rights, safety, or access to essential services may be affected.
- We provide appropriate information about an AI feature's intended use and material limitations so users can understand and critically evaluate its output.
- We promote an appropriate level of AI literacy for staff and others operating AI systems on our behalf.
- We apply privacy, security, data-governance, provider-review, monitoring, and incident-handling controls appropriate to the AI system and its use.
The legal classification of an AI system depends on its intended purpose and the context in which it is deployed. Customers remain responsible for assessing requirements specific to their own use of iMothership. This statement describes our operating commitments and is not a blanket certification that every use case is compliant.
AI Service Compliance Matrix
| Service | Models available | Status | Default API Policy | Key Requirement / Action Needed | Provider policies |
|---|---|---|---|---|---|
| Anthropic |
| Compliant | No Training | Anthropic does not train models on API inputs or outputs by default. | |
| Google (Gemini / Vertex API) |
| Compliant | No Training | Paid Google Cloud and Gemini API requests do not train models. Avoid the free tier. | |
| OpenAI |
| Compliant | No Training | Standard API platform data has not been used for training by default since March 2023. | |
| Replicate |
| Compliant | No Training | API inputs and outputs are not used for global model training. | |
| Fireworks AI |
| Compliant | No Training | API data is kept private and is not used for model training. | |
| Groq |
| Compliant | No Training | API data is isolated and is not used for model training. | |
| Local (Self-Hosted) |
| Compliant | 100% Isolated | Full local or on-premises execution provides complete data isolation. | Controlled by iMothership's private infrastructure and security policies. |
| ElevenLabs |
| Conditional | Zero-data retention confirmation is required in-app to use this service | Voice API usage must be configured for no training. Before ElevenLabs may process data, iMothership requires a workspace administrator to confirm that “Improve the models for everyone” is disabled under Terms and privacy > Data use. | |
| SpaceX AI (xAI / Grok API) |
| Conditional | Zero-data retention confirmation is required in-app to use this service | Enterprise API usage does not train models, while web and consumer tiers may train. All xAI API requests must use a team with Zero Data Retention enabled. iMothership requires every successful API response to include x-zero-data-retention: true and rejects the response if this cannot be confirmed. |
“Compliant” describes our assessment of the provider's published API or business-service controls for iMothership's intended use. It is not a general certification of every provider product, plan, model or optional feature. Customers should review the linked policies for requirements specific to their use case.