Pondera turns curriculum into a navigable knowledge constellation.
A private Kâ12 mathematics system that maps prerequisite knowledge, interprets diagnostic evidence and selects the next useful learning step. A working fractions slice is now ready for testing in our home-education setting.
The learner-facing card is not the product boundary. The durable core is the connected map: skills, prerequisites, misconceptions, evidence, and review paths.
Same answer. Three very different systems.
A learner answers 3/8 to 1/4 + 2/4. Pondera treats that as a small diagnostic event: what did the learner probably believe, and what should be repaired next?
Incorrect.
The system only knows the answer was wrong. The next activity is guesswork or a generic retry.
More practice needed.
The system knows the broad topic is shaky, but it does not know which idea inside the topic broke.
Repair denominator meaning.
The graph points to a likely misconception: the learner counted shaded pieces but treated denominators as addable counts.
Ready students should not wait for the calendar.
A broad graph lets Pondera move quickly through secure territory, pause on shaky edges, and route around a precise misconception when a wrong answer reveals one. The path is not fixed by a workbook order.
Evidence arrives
A check records what a learner did, not just whether the answer was right.
Find the node
The event attaches to a knowledge point, prerequisite, or misconception.
Update state
Mastery, confidence, and likely blockers shift in the local graph.
Choose the next star
The next step can accelerate forward, review, or remediate.
Keep it inspectable
AI-assisted content is checked against graph facts and deterministic rules.
The interface can change. The evidence trail should survive.
Different clients may want different learner surfaces. The durable part is the model beneath them: a wrong answer becomes evidence, evidence updates the graph-local learner state, and the graph picks a reasoned next move.
domain: mathematics
local_area: fractions
prompt: 1/4 + 2/4
answer: 3/8
verdict: incorrect
signal: likely_added_denominators
next: denominator_meaning_repair
audit: inspectable
Structure first
Curriculum standards are useful, but Pondera's primary unit is the teachable, diagnosable knowledge node.
Errors have shape
A mistake can point to a misconception, a missing prerequisite, or a fragile representation.
AI stays bounded
Hints and wording can vary, but graph IDs, checker logic, and evidence rules stay validated.
Maths learning should move at the speed of actual understanding.
Pondera grew out of a problem I kept coming back to for years: how do you help a child move through maths at the speed of their actual understanding, without turning learning into either a worksheet conveyor belt or a black-box AI chat?
I have a maths background, and I am building this first for home-education use because I wanted something better for my own kids: a system that can see the structure underneath a mistake, explain why the next step makes sense, and keep the learning path inspectable.
What exists today: a schema-validated curriculum graph, graph explorer, diagnostic and evidence model, bounded AI-assisted authoring checks, and a focused fractions slice ready for learner testing. Broader curriculum coverage and product packaging remain in development.
Lachlan Bridges, Founder
Early alpha, ready to test a focused slice.
Pondera is currently a private internal tool. The fractions slice is ready for hands-on testing with my children; broader curriculum coverage, learner flows and product packaging remain active work. The long-term path may be a paid product, a sale, or an open-source release, so the implementation and curriculum data stay private while that direction develops.