AI literacy for every student.
A self-paced AI literacy course built into a platform — students learn how to use AI the right way, from practice, in real time. Like Duolingo, but for AI.
Lesson 3 · Practice
Writing Effective Prompts
Use what you've learned to write a prompt that gets the AI to explain a complex topic simply.
Built for schools · No credit card to pilot · Responds within 24 hours
How it works
The same 4-step loop,
every lesson.
Only step 3 calls the model — and it's capped at 3 attempts permanently. The system is the lesson.
Read
The Lesson
Static text and examples written by the Velo team. No AI. No fluff. Just clear material the student reads at their own pace.
FreeUnderstand
The Concept
2–4 multiple-choice questions. Instant feedback. The goal is comprehension, not a grade. No AI cost at this step.
FreePractice
With Real AI
The student applies the concept using a live AI model. Exactly 2 attempts — no more, ever. The only step that calls the LLM.
2 attempts maxUnlock
Next Lesson
On completion the next lesson unlocks. New monthly lessons release on the 1st — keeping the school subscribed.
Auto-advanceWhy exactly 2 attempts?
Attempt 1 is to try. Attempt 2 is to apply what you learned. More than 2 is waste — both pedagogically and financially. The cap is permanent, enforced server-side before any API call.
Content schedule
30 lessons. Drip-fed
by design.
Month 1 hooks the students and proves value to the school. Month 2+ keeps them returning — and the school subscribed.
Month 1
Foundation Course
Month 2+
Ongoing — 10/month
Why the content gate exists
Content gates create predictable MRR. Students cannot binge all 30 lessons in a weekend — the platform enforces a monthly release schedule tied to the school's subscription start date.
Built-in guardrails
Limits that protect
everyone.
Every limit exists to protect unit economics. None are optional. All enforced before the LLM is ever called.
2 attempts per lesson — ever
After 2 attempts the practice step is locked permanently. Enforced server-side before any API call. Exceeding the limit shows a blocked UI, not an error.
40 students per cohort
Each school cohort is capped at 40 active students. Schools with multiple year groups pay per cohort — a natural upsell without limiting access.
400 token cap per response
Every LLM call is hard-capped at 400 output tokens. Keeps responses focused, prevents abuse, and makes per-student API cost fully predictable.
Monthly content gate
Next month's lessons do not unlock until the 1st of the next calendar month. Students cannot binge ahead. Checked in the gate layer — never trusted to the client.
Blocked state, not a crash. When any limit is hit, the student sees a clear, friendly message — never an error page. You can always advance without the practice step if your attempts are exhausted.
Get started
Run a pilot with
your school.
We'll set up a cohort, onboard your first group of students, and help you see results before you commit to a full plan.
Request PilotNo commitment · We'll be in touch within 24 hours