Make the deployment auditable before the claims get bigger.
AI grading needs evidence before claims scale. Planisphere records the review around tutor and grader workflows: gate the boundary action, route a named human, seal the record — signed evidence for Title VI, FERPA, and compliance review, with student records kept out of the evidence path.
The record crosses the boundary. The student record never does.
A grader scores. A tutor answers. The district may later need to show what happened, who reviewed it, and whether student records stayed protected. Planisphere seals one tamper-evident record per gated tutor or grader action — hashes, decisions, timestamps, and the named human who signed off — and that record crosses the boundary without carrying a student’s work. When a parent, a board member, or the Office for Civil Rights asks how the district supervised its AI graders, this is the record you hand them.
The education Toolkit: five primitives, pre-arranged.
The education Toolkit is the same five primitives — gate / route / seal / verify / meter — pre-arranged for student-data boundaries and scoring-evidence duties under FERPA, COPPA, and IDEA. It is a catalog-backed pack bundle on the hosted API: record verbs plus the regulation catalog. The full reference map lives at the education library surface. See how the five primitives compose, or read the pack entry in the live registry.