Abyssale offers powerful AI capabilities to enhance your design workflow. As a workspace administrator, you have control over how these features are implemented across your team.
What are AI Features and Content Training?
Abyssale's AI features help streamline your design production process through automation and intelligent assistance. Content training allows Abyssale to improve its AI models by learning from anonymized workspace data, helping the system better understand design concepts and patterns.
Managing AI Settings
Who Can Access These Settings
These AI management settings are available to workspace owner & administrators across all subscription plans. The settings can be found in your workspace settings area.
Two Key Controls
Administrators have control over two primary AI settings:
AI Features Access
Controls whether team members can use AI features within the workspace
Set to 'ON' by default for all plans
AI Model Training
Controls whether your workspace data can be used (in anonymized form) to improve Abyssale's AI models
Default settings vary by plan:
For Self service Subscription plans: 'ON' by default (admins can opt out)
For Enterprise plans: 'OFF' by default (admins can opt in)
How to Adjust AI Settings
Navigate to your workspace AI settings
Use the toggles to enable or disable:
AI Features Access: When ON, users can access AI features; when OFF, AI features are disabled for all users in the workspace
AI Model Training: When ON, anonymized workspace data may be used to improve AI models; when OFF, your data will not be used for training
Effects of Your Settings
AI Features Access
When enabled: Team members can use all AI features available in their subscription plan
When disabled: AI features will be unavailable to all users in your workspace
AI Content Training
When enabled: Anonymized data from your workspace may help improve Abyssale's AI capabilities
When disabled: Your workspace data will not be used to train AI models
Changes to these settings take effect immediately.
Controlling Individual AI Models
In addition to turning AI features on or off workspace-wide, administrators can control which specific AI models their team is allowed to use, and which one is used by default.
This is useful if you want your team to only use certain models, for example, to standardize output quality, control credit consumption, or restrict usage of models that don't meet your organization's content or compliance requirements.
Where to find it
Go to your workspace AI settings and from the Models preferences section, you'll see a list of models grouped by feature (Text to image, Remove Background, etc.), along with:
The model's provider (Google Gemini, OpenAI, Alibaba Cloud, etc.)
Supported resolution/aspect ratio tags
Credit cost per generation
A toggle to enable or disable the model
Enabling or disabling a model
Use the toggle next to any model to turn it on or off:
ON (green): The model is available for your team to select when generating content
OFF (gray): The model is hidden from selection, team members will not be able to use it, even if it's included in your subscription plan
This lets you restrict access to specific models across your workspace — for example, disabling models you don't want used for cost, quality, or compliance reasons.
Setting a default model
Click the ⋯ (more options) button next to any enabled model and select Set model as default. This model will then be pre-selected whenever a team member starts a new AI generation, making it easy to standardize on a preferred model without relying on each user to choose the right one manually.
A model must be enabled before it can be set as the default. Disabling the current default model will require selecting a new default.
Why this matters
Combining these two controls lets administrators:
Prevent use of models that don't meet internal content, quality, or compliance standards
Ensure the whole team defaults to a consistent, approved model
Manage credit consumption by disabling higher-cost models when not needed
Data Privacy and Security
Abyssale prioritizes the security and privacy of your data. Our model development process is designed to protect your confidential information:
All data used for training is anonymized
Sensitive information is redacted from text and images
We focus on learning general design patterns and Abyssale-specific concepts—not your specific content or ideas
The AI training process is designed to improve functionality without compromising your proprietary designs or information.