This kenzoAI Data Policy explains what happens to the information you submit when you use kenzoAI within BuildNinja or use Dojo. It supplements, and should be read together with, the BuildNinja EULA, the kenzoAI Terms, and our general Privacy Policy. Capitalized terms not defined here have the meaning given in the kenzoAI Terms.
This policy applies only to kenzoAI and Dojo. Core BuildNinja remains self-hosted on your own infrastructure, as described in EULA Section 9.6, and is not affected by anything in this policy.
1. Data Processed by Kenzoai
When you submit a kenzoAI request from a supported BuildNinja workflow or another AI-assisted feature, or when you submit a request through Dojo, we process the data you include, or BuildNinja includes, with that request (such as build information, logs, source-code or configuration diffs, and configuration data), together with metadata needed to fulfil and meter it (timestamp, selected Model Tier, requesting account, and kenzoAI Credit deduction). kenzoAI is not intended for, and you should not submit, Sensitive Information (as defined in our Privacy Policy) or any data you do not have the right to submit.
Account and plan data. Separately from request content, we process, as controller, the account data we use to provide and meter kenzoAI: your internal user ID and plan details (start and end dates, credits, the Model Tiers allowed, the duration for which activity logs and app usage data are stored, and the maximum number of apps), received from GrapeHub MyAccount, and kenzoAI Credit deductions. We do not send your name, email address or phone number to the Developer Portal.
2. Request Handling
A kenzoAI request is sent to one of our AI Providers solely to generate the response to that request. We currently use Microsoft Azure AI Foundry and AWS Bedrock as our AI Providers. Which provider serves a given request depends on your selected Model Tier and our internal routing; we may change this routing over time without changing this policy.
kenzoAI processes the submitted data and uses models hosted by our AI Providers to generate and deliver AI Output as part of the service.
3. Model Training and Retention
3.1 Provider Training Practices
Based on each AI Provider's current published terms, your submitted content is not used to train that provider's models, and is not shared with unrelated third parties. Each AI Provider does, however, operate a limited exception for abuse detection, described in Section 3.2 below.
3.2 Abuse-Detection Exception
- Microsoft Azure maintains an abuse-monitoring data store where prompts and completions may be retained under Microsoft's published Azure OpenAI Service abuse-monitoring terms. If flagged by automated monitoring or a suspected abuse pattern, a limited set of authorized human reviewers may review them.
- AWS Bedrock is configured to the default data retention mode, rather than none or provider_data_share, for requests routed to it.
3.3 Database Retention
We maintain internal request-response logs containing the content of kenzoAI requests and the corresponding AI Output. Separately, we store the content of your prompts and AI Output in our database for session management; this data is consequently included in database backups. Database backups are encrypted.
3.4 Changes to Provider Terms
This retention posture is governed in part by each AI Provider's own terms, which are outside our direct control and may change. If you are submitting content subject to specific regulatory or contractual retention requirements, consult this policy and the applicable AI Provider's current terms before doing so, or contact us first.
3.5 Other Retained Data
Separately from AI Provider retention and internal request-response logging, we retain the following:
- Request metadata (kenzoAI Credit deductions, timestamps, selected Model Tier): retained in accordance with the audit-log retention period of the plan you have purchased.
- Billing records: retention is governed by our subscription/billing team's own policies, separate from this kenzoAI service.
- Support records: no separate retention period is defined for kenzoAI support communications beyond whatever retention policy governs the mailbox or mail server involved.
- Developer Portal records (internal user ID and plan details): removed after the plan has ended, as per the plan's audit-log duration, and erasure is scheduled for 30 days after we receive your erasure request.
4. Dojo
Requests made through Dojo, the shared no-signup sandbox, are handled the same way as metered kenzoAI requests with respect to the AI Provider and retention posture described above, except that Dojo requests do not deduct kenzoAI Credits and are instead subject to the global shared token pool described in the kenzoAI Terms.
5. User Choices
You choose, per request or by app-level default, whether to use kenzoAI at all, and which Model Tier to use. If you do not want data related to your BuildNinja workflows processed by an AI Provider, do not submit a kenzoAI request; core BuildNinja functions independently of kenzoAI.
Erasure requests under this policy cover your kenzoAI data only, including your kenzoAI Developer Portal entitlements. We do not erase your GrapeHub MyAccount data as part of such a request, and erasing your GrapeHub MyAccount data does not erase your kenzoAI data. To erase your GrapeHub MyAccount data, you need to request that separately from GrapeHub MyAccount.
To ask a question about this policy, or to request more information about how a specific kenzoAI request was handled, contact our Data Protection Officer at dpo.india@grapecity.com.
6. Changes to This Policy
If we change how kenzoAI processes your data, including a material change to an AI Provider's retention or training practices that we become aware of, we will update this policy and take reasonable steps to bring the change to your attention, consistent with the notice approach described in our Privacy Policy.