- Category
- AI-personalized K–12 learning platform
- Audience
- Homeschool and after-school families
- Status
- Release preview, Fall 2026
A curriculum that bends without breaking
OHM is an AI-personalized K-12 learning companion for homeschool and after-school families, shaping each lesson around what a child already cares about without letting the interest replace the standard underneath it.
Personalization and structure are usually sold as opposites. A structured curriculum is legible to a parent but identical for every child; a personalized one is engaging and impossible to audit. OHM holds both at once: a K-12 pathway with mastery checkpoints and standards mappings, and an engine that decides how each concept gets taught from what the learner is actually interested in.
That shows up most clearly in the contextual layer. A child obsessed with fashion meets geometry through garment cuts; a skater meets quadratics through the arc of an ollie. The interest is the vehicle the concept arrives in rather than a reward bolted onto the end of a worksheet, and the objective underneath never moves. Every lesson opens three ways — video, visual diagram, or readable guide — with the AI mentor holding the thread across whichever the student picks.
The product is equally deliberate about what it is not. OHM states in the hero, the academic section, and the FAQ that it is a supplemental learning platform rather than an accredited school. Standards mappings are proposed for review instead of asserted, illustrative scenarios are labelled as illustrative, and the mentor ships with bounded fallbacks for when a model provider is unavailable.
Why personalized learning keeps failing families
The factory model was built for a different century
Lectures, cramming, and a bell schedule designed around shift work still set the shape of a school day. None of it reflects how the brain actually consolidates knowledge, and a child who falls out of step with the timetable simply falls behind it.
Personalization usually means an unauditable black box
Adaptive platforms adjust in ways a parent cannot see or check. When the path is invisible, a family has no way to tell whether a child is being challenged, coasting, or quietly routed around a gap.
Interest is treated as a bribe, not a route
Most products use a child's obsessions as a reward after the real work — badges, unlockables, themed skins. The subject matter itself stays generic, so the thing the child cares about never carries the concept.
Parents get dashboards, not answers
Progress tooling tends to produce charts that require interpretation. A guardian wants to know what was mastered, what needs another look, and what the child chose for themselves — in about five minutes.
AI claims outrun what can be evidenced
An AI tutor that promises a standards-aligned course and produces an unreviewed one puts the family's compliance risk ahead of the vendor's. The mapping needs a review gate, and the platform needs to keep working when the model does not.
One format for every learner
A lesson delivered only as video loses the student who reads, and a lesson delivered only as text loses the student who needs to see the parts move. Forcing one format through a whole curriculum guarantees a mismatch somewhere.
How we built it
Anchor to a structured K-12 pathway first
The curriculum map comes before the personalization. Core subjects, electives, mastery checkpoints, and a progress record exist independently of whatever route the engine composes through them, so there is always something for a parent to check the path against.
Make interest the vehicle, not the reward
Contextual learning teaches the same objective through the thing the learner is already thinking about — garment cuts for geometry, ollie arcs for quadratics, ramp angles for physics. The concept is fixed; only its carrier changes.
Run the lesson as a four-move loop
Sense the learner's strengths, gaps, pace, and preferences. Compose one connected lesson from them. Respond in real time when the student stalls or accelerates. Prove mastery through work they produce. Then repeat, sharper.
Open every lesson three ways
Video, visual, and text are three doors into the same idea rather than three different lessons. The AI mentor holds the thread across whichever the student picks, so switching format does not restart the concept.
Build the forge, then gate it
The Course Forge drafts a full multi-week course from a single named obsession — octopus intelligence, the economics of K-pop, skateboard physics — with a capstone and proposed standards mappings. Those mappings are explicitly proposed for review, not published as approved.
Write the guardian view as prose
The weekly digest reads in five minutes and answers three questions: what was mastered, what is worth a second look, and what the child chose themselves. The full dashboard sits behind it for anyone who wants the detail.
Design the honest fallbacks in from the start
The AI mentor carries bounded fallbacks for provider outages, illustrative scenarios are labelled as illustrative rather than passed off as student testimonials, and the platform states its non-accredited status wherever a family might otherwise assume otherwise.
Capabilities
Dynamic pacing
Mastery accelerates the path; a gap reroutes it, reteaching the same principle through a different lens rather than repeating the failed explanation.
Contextual learning
Each concept is taught through an interest the learner already holds, so the subject arrives inside something they were going to think about anyway.
On-demand AI mentor
A lesson-aware tutor that re-explains ideas and generates fresh practice, with bounded fallback guidance when a model provider is unavailable.
Three ways into one lesson
Every lesson opens as a short video explainer, a visual diagram with guided observation, or a readable guide with vocabulary and step-by-step notes.
The Course Forge
Names an obsession and drafts a multi-week course around it — lessons, a capstone project, and proposed standards mappings held behind a mentor-review gate.
Multilingual by default
Learn in a native language while building English fluency, with translated explanations, bilingual vocabulary, pronunciation coaching, and bilingual modes.
Spaced review scheduling
A review schedule built from prior attempts and review outcomes brings concepts back over time rather than leaving them behind after one pass.
Active retrieval and explain-back
Comprehension is verified through practice, whiteboard work, and explaining concepts back — production rather than passive watching.
Metacognitive logs
Students track their own focus and cognitive peaks, making learning to learn a subject in its own right rather than a by-product.
Project portfolio with evidence
A submitted project is connected to lesson objectives and retained as evidence, with publication review preventing an unsupported standards claim from being treated as approved.
Guardian dashboard and weekly digest
A five-minute digest covering what was mastered, what needs a second look, and what the child chose — with full detail in the dashboard behind it.
Core, elective, and on-demand curriculum
Mathematics, English, Science, Social Studies and Health & PE as core; Arts Studio, Tech & AI, Languages and Entrepreneurship as electives; anything else spun up on request.
The visual language
OHM is dressed like a well-made book rather than a classroom app. An Instrument Serif display over a clean sans body, a near-white paper ground, and a deep teal primary keep the register calm and adult — the parent is being addressed as seriously as the child. Colour does the subject-tagging work so the typography never has to shout.
Colors
Deep Teal
#1A5F7A
Primary actions, arts tagging, and headline emphasis
Plum Teal
#00475E
Deepest ground, marine tagging, and high-contrast panels
Growth
#159895
Mastery states, entrepreneurship tagging, and forward motion
Curiosity
#B0791F
Progress crest and the markers for learner-chosen work
Electric
#F2921D
The single high-energy accent, used sparingly on calls to action
Knowledge
#2F7899
Secondary data and supporting interface tint
Paper
#F7F9FB
Base field, kept near-white so lesson content carries the contrast
Ink
#191C1E
Primary reading copy and display type
Muted Ink
#40484D
Secondary copy, captions, and metadata
Rule
#D8E1E6
Card borders, dividers, and table rules
Typography
Instrument Serif
Sets headlines and section titles, with the emphasised word in italic. A book serif rather than a rounded app face, because the product is asking to be trusted with a child's education.
Instrument Sans
Carries body copy, lesson text, controls, and the guardian digest. Sized for long-form reading rather than dashboard scanning.
Monospace stack
Eyebrows, subject tags, grade bands, and status strips. The monospace register marks the parts of the interface that are structural facts rather than prose.
Principles
- 01
Structure is what makes personalization safe
The K-12 pathway, mastery checkpoints and standards mappings exist so that an adaptive route always has something auditable to be measured against.
- 02
Interest carries the concept, it does not replace it
Contextual learning changes the vehicle, never the objective. The geometry a fashion-obsessed student learns is the same geometry.
- 03
Say what is proposed and what is proven
Standards mappings are proposed for review, scenarios are labelled illustrative, and the non-accredited status is stated wherever a family might assume otherwise.
- 04
The parent view is prose, not telemetry
A five-minute digest that reads like a conversation beats a dashboard that requires interpretation. Detail is available, but the summary is the product.
- 05
Degrade honestly
When a model provider is unavailable the mentor falls back to bounded guidance rather than failing silently or inventing an answer.
Questions
OHM is an AI-personalized K–12 learning companion for homeschool and after-school families. It combines a structured curriculum, adaptive practice, an on-demand AI mentor, and progress reporting a parent can read in a few minutes.
No. OHM is a supplemental learning platform, not an accredited school or a diploma-granting institution. It provides structured lessons, mastery practice, progress visibility, and a standards-mapping workflow that passes through publication review.
Four moves run behind every session. OHM senses the learner's strengths, gaps, pace and preferences; composes one connected lesson from them; responds in real time when the student stalls or is ready to move ahead; and verifies mastery through work the student produces.
Yes. OHM is designed to support learning alongside a traditional school, a homeschool plan, or family-led study. Lessons can reference standards mappings, but those mappings and the assessment evidence remain subject to the publication-review gate.
Less than homework help. The guardian dashboard shows what a child learned, struggled with, and chose for themselves, and a weekly digest lands in the inbox. No parent has to teach a lesson.
A generator that takes a single interest — octopus intelligence, the economics of K-pop, skateboard physics — and drafts a multi-week course around it, with lessons, a capstone project, and proposed standards mappings that a mentor reviews before they are treated as approved.
OHM does not currently run live clubs, debate nights, or regional meetups. Families using it as a supplement or a homeschool resource are expected to arrange in-person and group learning alongside the platform.
Any laptop or tablet with a browser and an internet connection. There is no special hardware requirement, and no install.
Interested in OHM School?
Tell us how you would like to be involved — investing, building, advising, or simply following along — and we will be in touch.
Register Your Interest
Work with Vital Ventures
We study, support, and build the foundational systems that shape how societies function. If you are working on something in that direction, we would like to hear about it.
