資料
Content Integration Overview

Content Integration Overview

Use Content Integration to expand existing customer-owned JSON and control TutorFlow-authored content from external systems.

Content Integration is the layer for connecting external content systems with TutorFlow content. It is designed for two flows:

  • Import existing source JSON, such as a customer's category, level, and lesson structure, and expand it into TutorFlow authoring outputs.
  • Control content created in TutorFlow so another system can retrieve, update, and reconcile resources.

Content Integration is not the Agent Platform API. Use /v1/content/** and tf_content_ keys for human-owned TutorFlow content that appears in classrooms and the admin UI. Use /v1/platform/** only for AI agent workflows that were explicitly built for the Agent Platform.

Content API keys use the TutorFlow organization's normal AI Credit balance. Standard resource CRUD is free when the caller supplies the content. AI-powered expansion outputs consume 3 credits for an interactive module, 1 credit for a summary video structure, and 3 credits for an expanded quiz. Check the available balance with GET /v1/content/credits before submitting expansion jobs.

Looking to surface classroom content inside an LMS instead of another content system? See the LMS Integration (LTI) section — it deep links courses, modules, slides, and tests into Canvas, Moodle, and other LMSs with single sign-on and grade passback.

What the Content API covers

The Content API reaches seven content types. Each has resource endpoints under /v1/content/classrooms/{classroomId}:

Content typeEndpointExpansion output
Modules/modulesinteractive_module
Courses/courses
Videos/videossummary_video
Slides/slides
Tests/testsexpanded_quiz
Games/games
Simulations/simulations

Games and simulations are not part of the source JSON expansion flow. They are built directly through their own generation endpoints rather than as an output of an expansion job. See Games over the Content API and Simulations over the Content API.

A simulation is a model a learner perturbs rather than something with a score, so its generation endpoints mirror a game's while its brief describes quantities and controls instead of a loop and levels.

  1. Create a test Content Integration API key as a signed-in TutorFlow admin.
  2. Check the organization's available AI Credit balance with GET /v1/content/credits.
  3. Create a Content Integration job with one small source JSON payload.
  4. Store the returned job id in the external system.
  5. Poll the job status endpoint until it is completed or failed.
  6. Retrieve the result manifest and store each output's credit amount and charge time.
  7. Store the result manifest and any generated resource references in the external system.
  8. Test direct resource reads and updates if TutorFlow-created content must sync back.
  9. Use webhooks for production status updates once polling is verified.

Output priorities

  1. interactive_module, primary output for app insertion.
  2. summary_video, secondary output. Video generation creates the editable video structure. MP4 rendering is triggered separately after review.
  3. expanded_quiz, tertiary output for extra assessment coverage.