Inside the Essence Credit economy

Essence Credits are the single unit that prices every AI-metered feature on the platform — AI Tutor sessions, syllabus generation, lecture and avatar-video generation, real-time BCI focus-session pacing — while also being the currency tutors are paid in and the reward referrers earn. One unit, a lot of very different jobs. This piece walks through how a credit actually moves through the system, and why the underlying rate still needs validating against real compute cost.

One credit, one rate

The platform prices Essence at a flat $0.01 per credit. That rate is what turns a $5 Starter pack purchase into 500 credits, and it's the same rate every metered AI feature draws against, whether that's a quick AI Tutor exchange or a full lecture-video render. Keeping one flat rate rather than a different price per feature is a deliberate simplification: it means a learner only has to reason about one number, not a menu of per-feature prices that would drift independently.

Two pools, drawn in a specific order

Not every credit a user has access to comes from a wallet they've topped up. Every account also gets a real daily and weekly usage allowance, sized against what each subscription tier actually pays for, and that allowance is drawn down first — before the purchasable essence_balance wallet is ever touched. Explorer, Scholar, and Apex additionally get certain metered features, like syllabus generation, included free rather than charged per use at all, on top of their allowance.

Practically, that means most day-to-day usage on a paid plan never touches purchased credits — the wallet exists mainly for bursts above the included allowance, or for free-tier users who want to go beyond what their daily/weekly allowance covers.

VR mode doesn't cost anything extra

Worth calling out explicitly because it surprises people: VR mode is not a separately metered feature. It's a viewing mode for a lesson you're already enrolled in and have already paid for (or unlocked via your plan), not an additional charge layered on top. The credit system exists to price genuinely marginal compute cost — model inference, video rendering — not to monetize every surface of the product.

Credits as a reward, not just a purchase

Essence flows the other direction too. A member's first logged contribution to a collaborative project earns a flat 500-credit reward — real, immediate, and independent of anything a subscription tier gates. Referrals work similarly: a successful referral pays out in credits rather than cash, which keeps the reward inside the same economy it's meant to grow rather than creating a second, separate incentive system to track.

Where a credit actually goes when you spend one

Trace a single credit from purchase to redemption and the path looks like this:

• A learner buys a $5 Starter pack, crediting 500 Essence to their essence_balance.
• They open an AI Tutor session in the "socratic" persona; the session is metered per exchange, drawing from their daily allowance first if any remains.
• Once the daily allowance is exhausted, further exchanges draw from the purchased wallet balance instead, decrementing essence_balance directly.
• On the backend, that same session is served by a real model call — GPT for the interactive, low-latency tutor chat, Claude for longer structured generation like a syllabus or a career roadmap — so the credit charged is standing in for a real, metered API cost on the other end, not an arbitrary number.

The open question this system hasn't answered yet

None of the above is fabricated or simulated — the packs are real purchases, the AI calls are real model calls, and the allowance math runs against real usage. What hasn't been fully validated yet is whether the $0.01/credit rate and the allowance sizes actually cover the real cost of the AI compute they're metering at scale, across every plan tier, once render-heavy features like avatar video are used heavily rather than occasionally. That's an open unit-economics question the team tracks internally rather than one that's been quietly assumed away — worth watching if you're curious how sustainable a credit-metered AI product actually is once real usage patterns emerge.