Frequently asked.

The cap is $18.0M post-money. The instrument is a post-money SAFE. The cap is the same on every check in this round regardless of check size, and it does not move between now and close. The $18M anchor is triangulated by an IPEV study (six methods, weighted), which lives in the data room.

20% discount to the next priced round. MFN on subsequent SAFEs this round means if we issue a SAFE at a lower cap or higher discount to a later check, your economics are automatically adjusted to match. The MFN closes when the round closes.

Yes. The standard minimum is $25,000. We accept smaller checks ($10K floor) on a relationship basis where we have history. Pro-rata rights are granted via side letter on checks of $250,000 and above, which is the threshold that maps to the cap-table admin cost of tracking pro-rata.

The SAFE template lives in the data room (legal folder, Tier 2). Sign via DocuSign. Wire instructions are issued once the countersigned SAFE is returned. We co-sign within one business day of receiving funds and provide a confirmation packet (countersigned SAFE + cap-table entry + welcome note). The whole loop typically closes inside a week.

Rolling close. The round opened February 2026. Target final close is August 31, 2026. We close individual SAFEs as they arrive rather than batching. First funds were deployed the same week the round opened. The target is $3.0M aggregate by the final close date.

Yes. Two common patterns. One, a soft commit now with a wire in thirty to sixty days, which we document via a simple commitment letter that holds the cap. Two, a tranched wire (for example 50% now, 50% on a named milestone). Tranched wires are documented per tranche; the cap on later tranches is the same as the first.

Parent-paid subscriptions. A Builder plan at $149 a month is the default. A Family plan at $249 a month serves households with multiple kids. A school tier (Workshop) at $79 per kid per month opens in Stage 2. There is no advertising, no data resale, no freemium-to-paid funnel that sells kid attention.

The $149 monthly Builder plan is benchmarked to Juni Learning ($250+), Code Ninjas ($200+), and premium tutoring comps, not to Scratch (free) or Khan ($44 a year). The parent audience that buys kid coding at the premium tier is real and proven. The $149 tier is a floor we expect to raise toward $199 after the A/B test in Stage 3. The Founding-100 cohort locks at $99 forever as a gratitude and reference-case mechanism, not as a pricing signal.

Blended ARPU is $149 per paying household per month at the list-price anchor. The mix carrying that number is roughly 72% Builder monthly, 20% Builder annual (paid upfront at a 25% discount, $112 per month equivalent), and 8% Family plan. ARPU holds through the forecast because the Family plan upside offsets annual-prepay discount erosion.

Workshop (schools) is the Stage 2 motion. Priced at $79 per kid per month on a classroom seat minimum. Sales are driven by pilot referrals from parents who teach or administer, and by direct outbound to curriculum leads at districts already running coding programs. Workshop is not on the Y1 revenue path; it anchors Y2 onward.

Base case at Y5: blended ARPU $149, gross margin 68%, LTV $1,264, LTV:CAC 42x, monthly churn 8%, annual retention 85%. The high LTV:CAC reflects the organic-first GTM (viral badge loop, school pilot referrals, founder network) and the fact that paid acquisition activates only once the organic loop is proven. The financials page defends every line.

Scratch and Tynker teach block-based syntax without ever shipping a real app. Khanmigo is a tutoring chatbot layered on a worksheet. Kidgeni is image generation framed as creativity. Tekku is the first kid surface where a kid plans a real app, writes real code with a glass-box AI, and ships a live URL in under half an hour. The competitive page carries the full teardown.

Small real apps. A clap counter, a multiplayer trivia round, a daily-habit tracker with a badge, a grandma-weather dashboard. Each runs at a real URL on a real domain the kid can text to anyone with a phone. The ship is the product. Every session ends either with a shipped app or with an explicit save-and-continue so the next session lands back in the same build.

Typical sessions run 20 to 35 minutes of kid-time-on-task. The product soft-caps at turn 15 with a "want to take a break?" prompt and hard-caps at turn 30 with a "time to show your grownup" screen. The caps exist to keep the product a supplement, not a replacement for the rest of a kid's day.

Ages 8 to 14 is the core. The product is tuned around an 8-year-old reader at the low end and a 14-year-old builder at the high end. Under 8 works for paired sessions with a parent, but the independent-use threshold is 8. Above 14 the product still functions but is not our audience focus.

Every kid input passes through OpenAI Moderation first. The Tekku tuning layer is a classifier that recognizes kid-normal build-talk ("I want to kill the timer when the button is pressed") and removes false positives from the moderation queue before they reach the pedagogy model. Real kid content that should be escalated still escalates. The parent surface includes an incident log.

The FTC COPPA modernization rule takes effect April 22, 2026. Penalty is $51,744 per violation per day. Tekku is built to the modernized rule from day one. Parent-verified consent, 90-day default data retention, no training on kid data, explicit parent-initiated delete, and a no-training clause with every third-party model provider. The safety-model page carries the full posture.

Three layers. One, OpenAI Moderation on every input and every output. Two, a Tekku-tuned classifier that removes kid-normal build-talk from the false-positive pile. Three, a content-policy system prompt that constrains the pedagogy model to reject off-task prompts regardless of phrasing. The workspace is an app-building sandbox, not a free-form chat, which narrows the jailbreak surface structurally.

No. Every third-party model provider (Anthropic, OpenAI) is contracted with a no-training clause for Tekku traffic. Tekku itself does not train on kid data. Aggregate, anonymized concept-label data (not the kid transcript, the concept the session demonstrated) flows into the internal concept graph, which is the data asset we describe as the moat. Parents can opt out of aggregate contribution at any time.

Persona-verified consent at onboarding. Parent proves identity with a government ID scan (processed by Persona, deleted after verification). The parent then grants explicit consent for the kid account, including a readable summary of what data is collected, how long it is kept, and what the kid will experience. Consent is revocable one-click from the parent dashboard.

TODO(validation): exact current active household count is not yet published. The founding cohort target is 100 paying households by end of Stage 1. The validation page carries cohort detail as it clears for public display.

TODO(validation): the full cohort retention curve has not yet run past month four and is not yet cleared for public publication. Base-case model assumption is 8% monthly churn (85% annual retention). The retention signals section on the validation page will carry the curve when the month-six cohort lands.

TODO(validation): parent NPS has not yet been sampled at a statistically meaningful cohort size. Founding-100 cohort survey is scheduled for month three of Stage 1. The validation page will carry the result.

TODO(validation): the waitlist number is live in the admin dashboard but has not yet been cleared for external publication. Growth to date is driven by founder network, a handful of parenting newsletters, and organic referrals from the founding cohort. The gtm page carries the channel mix.

Yes, inbound from two district curriculum leads and one private-school network. All three are parked until Stage 2 because we will not deploy Workshop until the kid-facing retention curve is defended. School interest is a signal, not yet a pipeline.

One founder (CEO, user zero, product owner, daily driver) supported by a specialized agent team for orchestration, finance, narrative, creative doctrine, and platform. Seed funds four founding human hires in the first twelve months: two engineers, a head of trust, and a head of school GTM. The team page carries founder bio and the hiring plan.

Remote-first, distributed. No physical office planned before the Series A. The operating cadence is async weekly rhythms with a synchronous founder-led standup for the founding hires. The agent team runs 24 hours a day and is the reason the one-founder operating pattern works at this velocity.

TODO(validation): the advisor roster is being formalized. Current outreach covers kid-education (two former Outschool leaders), developer tools (a former Vercel product lead), safety and child psychology (one academic, one practitioner), and capital (an angel who co-led the prior seed). The team page carries the roster as each advisor signs.

Post-seed (including this extension) the founder holds the majority, the option pool sits at 15%, prior SAFEs and this round convert at the next priced round at their caps. Full cap table is in the data room under the legal folder (Tier 2). The counsel page summarizes the structure.

No. The agent team is the operating pattern pre-seed. Seed capital funds four founding human hires inside the first twelve months. The agent team stays in place for orchestration, finance, narrative, and creative doctrine. Humans do the work that requires human judgment; agents do the work that benefits from persistent context and tireless execution.

Base-case exit is a strategic acquisition at Year 5 in the $2B enterprise-value range. The outcomes page carries the IPEV triangulation and comparables. Early acquisition offers get evaluated on a scenario basis. We do not plan for an early exit, but we will not turn away a fair offer that returns a meaningful multiple to this round.

Three tiers. Strategic edu (Duolingo, Khan Academy, Code.org, Roblox, Common Sense Networks) with roughly 55% probability weighting. Education PE (GSV, Owl, Reach, Leeds) at 30%. Horizontal AI (Anthropic, OpenAI, Google, Meta) if one decides to own a vertical youth channel, at roughly 10%. IPO is the long-tail 15% case at the top of the plan.

Base case 5 to 7 years to the primary liquidity event. Secondary liquidity for seed investors may open earlier at the Series B or Series C if the round is oversubscribed; we will support, not orchestrate, secondary for aligned allocators. The outcomes page carries the exit-scenario ladder.