Your AI keys, data and limits
This page explains the operating controls around LMS Blueprint’s optional AI features: your provider key, the information sent with requests, usage estimates and the limits you set. It is for course owners evaluating data flows and ongoing costs. For feedback, Folio, tutoring and authoring capabilities themselves, read the separate AI product page.
AI, on your terms
Your key.
Your provider.
Your control.
Connect your own AI account. Requests go from your server to the provider you choose, with calls and estimated usage visible in LMS Blueprint.
Bring your own key changes the relationship.
- Choose your provider
- Use your own provider account and API key rather than a supplied shared key. Your provider’s model availability and account terms still apply.
- See the usage
- Review calls, tokens and estimated spend by feature and day. Estimates are a planning aid; the provider’s billing is authoritative.
- Set the boundaries
- Enable the AI features you want and configure usage controls. A provider account and model choice affect cost and availability.
- Pay for usage directly
- The AI add-on is a software subscription. Model usage is billed by the provider, separately from LMS Blueprint.
Local records. Explicit external processing.
Student records are stored in your WordPress database. When an AI feature is enabled, the information needed for that request is sent to your selected provider. “Stored on your server” does not mean an AI request stays inside it.
Payments, email delivery, video hosting and connected tools have their own processing boundaries too. Configure only the services you need.
Understand each connection →Three different jobs for AI.
- Feedback
- Respond to a student’s work in the context of your course and the feedback settings you define.
- Folio
- Keep work and relevant learning history connected so later interactions can use that context.
- Course tutor
- Help students with course-grounded responses and source references. AI responses still need appropriate teacher oversight.
Let an assistant help with the drafting.
The opt-in MCP connection supports authorised assistants such as your chosen AI assistant. Supported actions include reading course material and creating drafts. Permissions, limits and the action log belong in the setup, not as an afterthought.
Agent writes are drafts for review. A connected AI tool has its own access and processing rules; grant only the permissions needed.
Read the AI feature inventory →Plan with your own numbers.
A hosted allowance or credit bundle and a BYOK setup are different ways to buy usage. Compare the software, provider usage, hosting and support together, using the same workload.
Build your running-cost estimate →Inside the platform
A Recipe, with choices you own.
- The Academic Method Recipe
- Use a tested mix of models, updated monthly and signed and verified on your site. Accept updates automatically or with a click, or choose Best quality, Balanced, Value or a custom mix with estimated cost per student.
- AI budgets and a pause control
- Set a monthly money ceiling, receive usage alerts, pause all AI and test your key with a small request. Keys are encrypted and masked; AI calls go only to public addresses.
- Allowances shown before purchase
- Set per-student allowances and show what is included under buy buttons, at checkout and in the welcome email.
- Know what each AI request sends
- Inspect what is sent and what is never sent for each AI feature. Student-learning settings and course-creation settings have separate areas.
- Draft video questions and flashcards
- Draft question pauses from a video transcript or make a draft flashcard deck from a lesson, ready for your review.