The adaptive engine as an API.
Bring your own content in the unit-graph format and let the engine handle sequencing, learner modelling and variant testing. The API is in private preview with selected partners.
Six ideas, and you have the whole model.
Units
A unit is a few sentences with an optional check. It carries prerequisites, tags and one or more variants.
Graph
Units link into a directed graph. Prerequisites are explicit, so the engine can reason about what to teach next.
Sessions
A session walks the graph for one learner: request the next step, submit a response, repeat until the goal is met.
Learner model
Read the learner's estimated mastery per unit at any time, and seed it from prior knowledge you already hold.
Variants
Publish alternative phrasings of a unit. The engine tests them across sessions and reports the evidence.
Events
Webhooks for completion, struggle and milestone events, so your own systems can react.
Three calls to run a lesson.
Open a session, ask for the next step, send back the learner's response. The engine chooses among candidate units, updates the model, and tells you what comes next.
Illustrative. The preview API may differ, and access is granted per partner.
# open a session for one learner POST /v1/sessions { "course": "fin-acc-101", "learner": "lrn_8f2c" } # ask the engine what to teach next GET /v1/sessions/ses_31ab/next { "unit": "weighted-average-cost", "variant": "B", "candidates": 3, "prompt": "Batches: 100 at $10, 200 at $13, 100 at $16…", "choices": ["$12.50", "$13.00", "$14.00"] } # send the learner's response back POST /v1/sessions/ses_31ab/respond { "unit": "weighted-average-cost", "answer": "$13.00" } { "correct": true, "model": { "weighted-average-cost": 0.82, "fifo": 0.41 }, "next": "fifo" }
Tell us what you are building.
We onboard a small number of partners at a time and work closely with each. The more concrete your plan, the faster we can say yes.