Score B · 6.62/10Free to read$100K Hackathon · November 16, 2026 → December 16, 2026Nov 15, 2026

SpeakGrade Report Cards

SpeakGrade turns two short phone recordings of a child reading aloud into a logo-branded, parent-ready spoken-English report card that shows exactly which words and habits improved since the last term.

One product, one founder, thirty days. Everything below is built to be executed inside the contest window — the official rules still decide what counts.

Decide in 20 seconds

The facts that decide whether this is worth thirty days of yours.

Promising6.62/10
MVP build (solo)
4 days
Launch price
$19
Sales to $1,000 revenue
53 customers
First paying customer
Direct Instagram DM to independent English tutors who post student clips or advertise 1:1 kids' spoken-English classes.
Who pays

The owner-tutor of a 1-to-5-person online or neighbourhood English tuition business teaching children aged 6-14 in India, Pakistan, Vietnam, Indonesia or the Philippines, at the monthly or end-of-term parent renewal window.

Biggest execution risk

Operational simplicity 5.1/10

Why this could work in 30 days: one founder can ship the MVP in 4 days; the plan targets a first sale on day 10; 53 sales at $19 clears $1,000 of collected revenue; strongest signal is speed to first revenue at 8.2/10.

Score and projections come from the generated report and are not guarantees. The $100K Hackathon is run by Ship or Die — read the official rules before you build.

Your first 24 hours

12 concrete actions. Tick them off as you go — nothing is saved or sent.

0/12
Entry price
$19
Solo build
4 days
30d base revenue
$1,264
30d hackathon profit
$1,201
Section 1 of 8

Product opportunity and positioning

Target customer

The owner-tutor of a 1-to-5-person online or neighbourhood English tuition business teaching children aged 6-14 in India, Pakistan, Vietnam, Indonesia or the Philippines, at the monthly or end-of-term parent renewal window.

Why now

Managed speech-to-text with word-level timestamps is now cheap and accurate enough on child speech, so a single founder can ship a document-grade product in four days and sell it to micro-business owners who already record students on their phones.

Problem

At renewal time the tutor has no evidence of progress except memory, so parents ask what they paid for, haggle, or leave quietly.

Solution

The tutor uploads a 60-second reading clip per child; SpeakGrade transcribes with word timestamps, aligns it to the passage script, computes fluency metrics, diffs them against that child's previous report, and emails a one-page PDF carrying the tutor's own name and logo.

Differentiation · Competitors sell child-facing practice apps or generic language courses. SpeakGrade is a back-office artifact generator: no lessons, no AI character, no child logins, just the branded proof document the tutor's renewal depends on.

1. Product opportunity and positioning

Who pays. The buyer is the owner-tutor of a 1-to-5-person English tuition business teaching children aged 6-14 in India, Pakistan, Vietnam, Indonesia or the Philippines. They sell 1:1 spoken-English sessions at roughly $40-100 per child per month and invoice parents directly on WhatsApp. No procurement, no committee, no brand needed to reach them: a phone and a card are enough.

The trigger. Renewal. Every month or term end the tutor must answer "what exactly did my child learn for that money?" Today the answer is a voice note and "good progress!". Parents compare fees across tutors, haggle, or leave quietly. Writing individual progress notes for 25-40 students eats a weekend and still reads as opinion, so most tutors skip it and lose renewals without ever seeing the churn.

What SpeakGrade sells. Two 60-second recordings per child, taken a term apart, become a one-page PDF in the tutor's name and logo: a headline band, the words read correctly, substituted and skipped, words-per-minute, pause ratio, filler count, longest fluent stretch, a "words fixed since last report" list quoting the exact words, three practice drills and a next-term goal. The document belongs to the tutor, not to us. That matters, because the tutor hands it to a paying parent as their own professional work.

Why this wedge. The original product asked parents to buy a child-facing AI fox companion with a curriculum — a hard sell to a cold audience with no brand. SpeakGrade flips the buyer to the micro-business owner, shrinks the promise to one document, and lands on a payment moment that already exists every month. Protecting a $40-100/month tutoring fee with a $79 pack needs no ROI argument.

What JEV says. Weighted score 6.62/10, band B, confidence 0.73. The strongest criteria are speed to first revenue (8.23) and solo build feasibility (7.49); the plan leans on both. The weakest are monetization potential (5.34) and operational simplicity (5.14), which is why the plan sells one-time credit packs with no subscription promise and keeps every unit of manual work inside a two-minute review step. Market demand (6.3) and reachable customers (7.39) support direct outreach, not ads.

Positioning line. The report card a tutor hands a parent at renewal, generated in three minutes from a phone recording.

Section 2 of 8

Smallest sellable MVP

A mobile web app where a tutor uploads one 60-second reading clip, confirms the flagged words in a review screen, and downloads a one-page branded PDF report card — sold as a $19 five-report Trial Pack.

Must-have features

  • Student intake: name, level, one of 12 built-in levelled passages or a pasted custom script
  • In-browser phone recording plus audio file upload
  • Speech-to-text with word-level timestamps, cached per recording
  • Alignment engine producing correct / substituted / skipped word lists
  • Fluency metrics: words-per-minute, pause ratio, filler count, longest fluent stretch
  • Review-and-correct screen where the tutor taps to override flagged words
  • One-page branded PDF with logo upload, headline band, drills and next-term goal
  • Longitudinal delta block comparing against the same student's previous report
  • Credit-pack checkout and email plus WhatsApp-shareable link delivery

Explicitly out of scope

  • No child-facing AI voice companion or conversational practice
  • No curriculum, lesson plans, vocabulary courses or CEFR certification claims
  • No IELTS/PTE band prediction
  • No school, LMS or tuition-centre ERP integration, no multi-teacher seats
  • No native iOS or Android app in the first 30 days — mobile web only
  • No free-forever tier; only a paid trial pack

Demo flow (what the buyer sees)

  1. Tutor opens the app on a phone and adds a student with name and level.
  2. Tutor picks a levelled passage or pastes their own script.
  3. Tutor records the child reading for 60 seconds or uploads a voice note.
  4. SpeakGrade transcribes with word timestamps and aligns to the script.
  5. Tutor reviews flagged words and corrects any false positives in under two minutes.
  6. A branded one-page PDF is generated, emailed, and a share link is copied for WhatsApp.
Build plan

Technical implementation plan

Stack

Next.js on Vercel (mobile-first web)Supabase Postgres, Storage and AuthManaged speech-to-text API with word-level timestampsToken-diff alignment in TypeScript (word-level LCS)react-pdf or headless-Chrome HTML-to-PDFStripe Checkout for credit packs, connected to TrustMRRResend for PDF delivery email

Architecture

A single Next.js app on Vercel handles the mobile web UI and server routes. Supabase holds tutors, students, reports, credits and recordings; audio and logos live in Storage. On upload, a server route sends the clip to a managed speech API and caches the timestamped JSON against the recording so a re-render never pays twice. An alignment module diffs the transcript against the passage script into correct, substituted and skipped lists, then computes the four fluency metrics. Those metrics are snapshotted on the report row, and the delta engine compares the new snapshot with the student's previous report. A review page lets the tutor override flags before the PDF renderer produces the one-page artifact, which is stored, emailed via Resend, and exposed through a public share link. Stripe Checkout issues credits; credits decrement on generation.

Build steps

  1. Create the Next.js repo, deploy to Vercel, provision Supabase.
  2. Define tables: tutors, students, reports, report_metrics, credits, recordings.
  3. Seed 12 levelled passages (A1 to B1) as structured JSON.
  4. Build the intake form and MediaRecorder plus file-upload path to Storage.
  5. Integrate the speech API and cache timestamped transcripts.
  6. Implement the alignment diff and metric computation, unit-tested on three hand-checked clips.
  7. Build the review-and-correct table with per-word overrides.
  8. Build the PDF template with logo, headline band, three word lists, drills and goal.
  9. Add the delta block, credit decrement, Resend email and share link.
  10. Wire Stripe Checkout for $19/$79/$199 packs and connect the account to TrustMRR.

AI components

  • Managed speech-to-text with word-level timestamps for alignment and metric computation
  • Optional LLM call to draft the three practice drills and next-term goal from the substituted-word list, with a fixed template fallback
Hardest technical risk: Speech-to-text substitution errors on non-native child accents producing false 'mispronounced' words inside a document a parent reads, which is why the review screen is a hard requirement rather than a nicety.
Section 3 of 8

Pricing and monetization

Trial Pack

Entry offer
$19 one-time

For A tutor testing the report on 5 students before a renewal conversation

  • 5 report cards
  • Built-in levelled passages
  • PDF download and WhatsApp share link

Term Pack

$79 one-time

For An independent tutor with 15-40 students heading into a renewal month

  • 30 report cards
  • Tutor name and uploaded logo on every PDF
  • Longitudinal 'words fixed since last report' comparison
  • Practice-drill block per student

Centre Pack

$199 one-time

For A 2-5 teacher neighbourhood tuition centre reporting to parents as a business

  • 100 report cards
  • Centre logo and branding
  • Per-batch summary page
  • Priority email support

Model: One-time credit packs sold as PDF report cards, with volume tiers by tutor size and no subscription.

First revenue offer: Send the tutor a sample report card carrying their own name and logo, then offer their first 5 report cards free for current students if they send one 60-second clip today; ask for the $79 Term Pack once they have handed the first PDF to a parent.

Expected ACV: $79

Upsell path: Trial Pack at $19 converts to Term Pack at $79 the following month; Term Pack buyers move to Centre Pack at $199 when a second teacher joins; referral bonus of 10 report cards per introduced tutor adds packs without new acquisition spend.

Target customers
20
Revenue model
$1,580
Costs
$110
Profit model
$1,470
Section 4 of 8

First-customer acquisition

Primary channel

Direct Instagram DM to independent English tutors who post student clips or advertise 1:1 kids' spoken-English classes.

Where to find them

  • Instagram hashtags #onlinetutor, #spokenenglishforkids, #englishtutor filtered to Bangalore, Pune, Hyderabad
  • Google Maps listings for 'spoken English classes for kids' in Bangalore, Pune, Hyderabad
  • Google Maps listings in Lahore, Ho Chi Minh City and Jakarta
  • Instagram tutors posting reels of children reciting or reading aloud
  • Facebook groups for private tutors in India and the Philippines
  • Local speech-and-drama tutors on Instagram who already coach recitation

Exact outreach motions

  • Generate a sample report card with each tutor's own name and logo before sending anything.
  • DM with the sample PDF attached and the line: 'I made a sample progress report with your name on it — want the blank template for your students this month?'
  • Offer 5 free report cards for current students, no card required, returned within 24 hours of one voice clip.
  • Follow up on day 3 and day 7 with the sample PDF re-attached and one line about the renewal window.
  • Ask every paying tutor for one introduction to another tutor in exchange for 10 bonus report cards.

Message template

Hi [Name] — I saw your reel of the Grade 3 reading class. I built a tool that turns a 60-second phone recording of a child reading into a branded progress report card you can send parents at renewal: words read correctly, words skipped, words-per-minute, and what improved since last term. I already made one with your name on it — want me to send the PDF? First 5 for your current students are free.
Week-one goal: 150 tutors logged with handle, city and apparent student count; 60 personalised DMs sent with named sample PDFs; at least 6 replies and 3 tutors agreeing to send a first recording.
Section 5 of 8

30-day execution roadmap

MVP deadline: day 5First sale target: day 1030 days planned
Week 1 · Build & validate · 7 days
Day 1Foundation

Live payment path and app skeleton

  • Create Next.js repo, deploy to Vercel, provision Supabase.
  • Create tables for tutors, students, reports, credits, recordings.
  • Wire Stripe Checkout for the three packs and connect the account to TrustMRR.
  • Run a full checkout in test mode, then flip keys to live.

Deliverable: Live URL with three working checkout links and an empty dashboard.

Day 2Intake

Passages and audio capture

  • Write 12 levelled passages (A1 to B1) as seed JSON.
  • Build the form: student name, level, passage picker or paste-your-own.
  • Add MediaRecorder recording plus file upload into Supabase Storage.
  • Test on two real Android phones and one iPhone.

Deliverable: A recording lands in storage from a phone in under 60 seconds.

Day 3Analysis

Transcript, alignment, metrics

  • Send audio to the managed speech API with word timestamps and cache the JSON.
  • Align transcript to passage script into correct, substituted and skipped lists.
  • Compute WPM, pause ratio, filler count and longest fluent stretch.
  • Hand-verify metrics on three sample clips.

Deliverable: Analysis JSON for three clips that matches a manual count.

Day 4Output

Review screen and branded PDF

  • Build the review table where the tutor taps a word to mark correct or wrong.
  • Build the one-page PDF: headline band, three word lists, three drills, next-term goal.
  • Add logo upload and tutor name/contact block.
  • Render five PDFs and fix typography until it looks mailable.

Deliverable: A branded one-page PDF downloaded from a real recording.

Day 5Ship MVPMVP shipped

Delta, delivery, credits

  • Diff current metrics against the student's previous report for the 'words fixed' block.
  • Decrement credits on generation and block at zero with a buy link.
  • Email the PDF via Resend and create a WhatsApp-shareable link.
  • Publish a sample report card on the landing page.

Deliverable: MVP live end to end: upload, review, PDF, email, credits.

Day 6Proof assets

Sample cards and demo

  • Pick five public tutor profiles and generate sample cards with their names.
  • QA mobile web on iOS Safari and Android Chrome.
  • Fix the three worst visual defects.
  • Record a 90-second screen demo.

Deliverable: Five named sample PDFs plus one demo video.

Day 7List building

60 tutors logged

  • Collect 60 tutor profiles from the target hashtags.
  • Log handle, city, apparent student count and last post date.
  • Rank by posting recency and student volume.

Deliverable: A sheet of 60 qualified, ranked handles.

Week 2 · Launch & first sale · 7 days
Day 8List building

Reach 150 and add Maps

  • Add Google Maps listings for six cities.
  • Add Facebook tutor-group members who post classes.
  • Fill the sheet to 150 rows with a personalisation note per row.

Deliverable: 150-row outreach sheet with sample-card status per row.

Day 9Outreach

First 30 DMs

  • Generate a named sample PDF for each of the first 30 targets.
  • Send 30 personalised DMs with the PDF attached.
  • Log every send time for day-3 and day-7 follow-ups.

Deliverable: 30 DMs sent, 30 samples attached, reply tracking live.

Day 10OutreachFirst sale target

Next 30 DMs and first sale attempt

  • Send 30 more DMs with named samples.
  • Offer the free 5-report pack to every replier.
  • Pitch the $79 Term Pack to any tutor who asks about pricing.

Deliverable: 60 total DMs sent and the first paid pack attempted.

Day 11Delivery

Turn replies into recordings

  • Deliver every accepted free pack within 24 hours of the clip arriving.
  • Screen-record the PDF opening so the tutor can preview before paying.
  • Ask each tutor for permission to reuse the report as an anonymised sample.

Deliverable: At least 3 free packs delivered, 1 permission note on file.

Day 12Follow-up

Day-3 follow-up wave

  • Send the day-3 follow-up to all non-repliers with the PDF re-attached.
  • Ask repliers one question: which phrase do parents actually say at renewal.
  • Log objections verbatim in the sheet.

Deliverable: Objection log with at least 5 verbatim parent-related quotes.

Day 13Conversion

Term Pack close

  • Offer the $79 Term Pack with a 48-hour intro price to every delivered tutor.
  • Send a Stripe payment link inside the same WhatsApp thread.
  • Fix any checkout friction reported in the first 24 hours.

Deliverable: First paid Term Pack or a documented reason it stalled.

Day 14Review

Week 2 numbers

  • Count replies, free packs delivered, PDFs downloaded, packs paid.
  • Cut the weakest message angle and rewrite the DM opener.
  • Double the sample-card quality bar if visuals caused hesitation.

Deliverable: A one-page week-2 scoreboard with the winning DM variant.

Week 3 · Repeat & systematize · 7 days
Day 15Outreach

Volume to 100 DMs

  • Build 40 more sample cards from the sheet.
  • Send 40 DMs using only the winning opener.
  • Log reply rate by city.

Deliverable: 100 total DMs sent with reply rate per city.

Day 16Outreach

Centre outreach begins

  • List 25 neighbourhood tuition centres from Google Maps.
  • Prepare a Centre Pack one-pager with a per-batch summary example.
  • Send 25 WhatsApp messages to listed centre numbers.

Deliverable: 25 centre conversations opened.

Day 17Follow-up

Day-7 follow-up wave

  • Send the day-7 follow-up to all first-wave non-repliers.
  • Re-send any sample card the tutor never opened as a plain image.
  • Archive anyone who declined twice.

Deliverable: Follow-up wave complete, archive list cleaned.

Day 18Conversion

Second paid cohort

  • Push every delivered free pack toward the $79 Term Pack.
  • Offer 10 bonus cards for a same-week payment.
  • Confirm delivery cost per report is still under $0.30.

Deliverable: Second paid cohort of at least 3 tutors or a clear blocker.

Day 19Growth

Referral loop

  • Ask every paying tutor for one named introduction.
  • Deliver 10 bonus cards automatically on a successful introduction.
  • Add a referral field to the account page.

Deliverable: Referral field live and at least 2 introductions requested.

Day 20Product

Cut review time

  • Measure minutes spent per report on the last 10 reports.
  • Reduce flagged-word noise with a stricter pass threshold and a keep/undo toggle.
  • Add a one-click regenerate that reuses the cached transcript.

Deliverable: Median review time under two minutes on 10 reports.

Day 21Review

Week 3 numbers

  • Recompute reply rate, free-to-paid rate and average pack value.
  • Decide whether the Centre Pack or Term Pack is the real unit.
  • Rewrite next week's targets around the winner.

Deliverable: Week-3 scoreboard and a single chosen primary tier.

Week 4 · Scale what works · 7 days
Day 22Proof

Anonymised before/after page

  • Publish a before/after report pair with a paying tutor's written permission, name removed.
  • Add a three-line 'how the PDF gets made' block to the landing page.
  • Link the page in all follow-ups.

Deliverable: A public proof page with no invented names or metrics.

Day 23Outreach

Fresh 40 DMs

  • Add 40 new tutors from two untapped cities.
  • Send 40 DMs with the proof page linked.
  • Track which city replies fastest.

Deliverable: 140 total DMs and a per-city reply ranking.

Day 24Follow-up

Second follow-up on all waves

  • Follow up once on every unanswered thread older than three days.
  • Ask decliners one question about what would have made them say yes.
  • Update the DM opener with the top objection answer.

Deliverable: Objection-answer sentence tested in 20 sends.

Day 25Urgency

Term-end deadline push

  • Identify tutors whose schools are near term end and message them first.
  • Frame the offer against their parent-meeting date, not the price.
  • Offer the Centre Pack to any tutor with more than 40 students.

Deliverable: Term-end message variant sent to 30 tutors.

Day 26Conversion

Landing page fixes

  • Put the sample PDF above the fold and behind a one-field email capture.
  • Add pack comparison and a two-line refund stance.
  • Remove any claim the product cannot back with a rendered PDF.

Deliverable: Updated landing page with sample-PDF-first layout.

Day 27Growth

Deliver referral promises

  • Fulfil every promised bonus card.
  • Ask the two warmest tutors for a one-line public comment, name optional.
  • Add referral tracking to the credits table.

Deliverable: Referral bonuses delivered and at least one attributable introduction.

Day 28Review

Week 4 numbers

  • Recompute all funnel rates and compare to week 2.
  • Identify the one step that loses the most tutors.
  • Fix only that step before the final push.

Deliverable: Funnel table with the single biggest leak named.

Days 29-30 · Consolidate · 2 days
Day 29Close

Final conversion push

  • Message every warm but unpaid tutor with a hard end-of-window deadline.
  • Offer the Trial Pack at $19 to anyone hesitating on $79.
  • Answer every open thread the same day.

Deliverable: All warm threads closed, converted or marked dead.

Day 30Wrap

Verify and decide

  • Reconcile Stripe revenue against TrustMRR records and count refunds.
  • Confirm every sale was inside the contest window.
  • Write the next-30-day plan around whichever tier actually sold.

Deliverable: Reconciled revenue sheet and a written next-30-day plan.

Section 6 of 8

Revenue and profit scenarios

How the numbers are defined

Hackathon profit — Collected revenue − refunds − chargebacks − payment processing fees − paid marketing spend.

Business profit — Hackathon profit − operating costs (hosting, domains, tooling, AI/API delivery costs).

The contest ranks entrants on hackathon profit only. Hosting, tooling and AI/API delivery costs are not deducted by the organisers, but each sale still has to cover them — so we show both. Definitions verified on 2026-11-15 against the official rules.

Stated assumptions

  • Conservative/Base/Upside: every sale is a one-time credit pack, with the $79 Term Pack treated as the unit price.
  • Conservative: cold DM reply rate lands near 4% and paid conversion at roughly 10% of repliers, producing 4 packs.
  • Base: reply rate near 6% and paid conversion near 20% of repliers, producing 16 packs, in line with the concept's 20-customer hypothesis but discounted.
  • Upside: outreach reaches 200 tutors, referral adds 6 packs, and 30 packs close, with a small paid boost on Google Maps-local keywords.
  • Payment processing is assumed at Stripe rates, roughly 5% blended including cross-border card fees.
  • Refunds are set to zero in Conservative and Base because packs are prepaid credits; Upside carries one $79 refund.
  • Chargebacks are set to zero across all scenarios; digital credit packs with immediate delivery have no chargeback history yet.
  • Operating costs cover Vercel, Supabase, a domain, Resend, and $0.10-0.30 of speech-API cost per report delivered.
  • Hosting and AI delivery costs are excluded from hackathon profit per the contest rules but included in business profit.
  • All figures are projections from stated assumptions, not verified revenue, customers or traction.

Conservative

4 customers
$316
collected revenue · avg payment $79
Refunds
$0
Chargebacks
$0
Payment fees
$16
Paid marketing
$0
Hackathon profit
$300
Operating costs
$22
Business profit
$278

Base case

16 customers
$1,264
collected revenue · avg payment $79
Refunds
$0
Chargebacks
$0
Payment fees
$63
Paid marketing
$0
Hackathon profit
$1,201
Operating costs
$55
Business profit
$1,146

Optimistic

30 customers
$2,370
collected revenue · avg payment $79
Refunds
$79
Chargebacks
$0
Payment fees
$130
Paid marketing
$120
Hackathon profit
$2,041
Operating costs
$95
Business profit
$1,946

Model your own numbers

Runs in your browser · nothing is saved
Collected revenue
$304
Payment fees (3.9%)
$12
Paid marketing
$0
Hackathon profit
$292
Business profit
$237

Hypothetical arithmetic, not a forecast. The fee rate is a flat 3.9% assumption, refunds and chargebacks are left at zero here, and none of this is tracked data.

All scenarios are projections derived from the stated assumptions. They are not verified revenue, not existing customers, and not traction. Actual results will differ.

Section 7 of 8

Main risks and validation experiments

Risk 1Speech-to-text mislabels non-native child pronunciation and the PDF names a word the child actually said correctly, in front of a paying parent.

Mitigation

Mandatory review-and-correct screen before the PDF renders; flagged words are suggestions, never final text.

Validation experiment

Run 10 real child clips through the pipeline and count false flags per report; ask one tutor to time their correction pass.

Kill criteria

If median correction time exceeds five minutes, or false flags exceed 8 per report after threshold tuning, the document quality bar is unreachable and the wedge must change.

Risk 2Cold Instagram DM conversion sits far below the assumed paid-conversion rate, so 16 sales is unreachable.

Mitigation

Same-day rewrite of the opener around the top verbatim objection; add Maps and Facebook group channels.

Validation experiment

Send 30 DMs with the named sample and 30 with plain text; compare reply rates before scaling to 150.

Kill criteria

If 60 DMs produce fewer than 3 replies, the personalised-sample angle is wrong and the channel must be replaced, not tuned.

Risk 3Tutors accept the free pack but never send a recording, so nothing converts to paid.

Mitigation

Make the free path a single WhatsApp voice note, no app signup required for the first report.

Validation experiment

Offer the free pack to 10 tutors and count how many send a clip within 48 hours.

Kill criteria

If fewer than 3 of 10 acceptors send a clip, the effort barrier is too high and intake must move fully into WhatsApp.

Risk 4Read-aloud scoring measures reading fluency, not spontaneous speaking, and a sophisticated parent or competitor calls it out.

Mitigation

Label the artifact explicitly as reading fluency and pace, never as overall spoken-English ability or a CEFR level.

Validation experiment

Show the PDF to 5 tutors and ask whether a parent would challenge the framing.

Kill criteria

If tutors say they would not hand it to a parent as written, the framing must be reworked before further selling.

Risk 5One-time packs produce a spike, not a business, and month two collapses.

Mitigation

Time the Term Pack to each tutor's renewal calendar and introduce a repeat-purchase nudge at day 25 of their credit cycle.

Validation experiment

Ask every paying tutor for their next parent-meeting date and log it.

Kill criteria

If fewer than 3 of the first 10 buyers have a dated renewal event within 60 days, the recurring assumption is dead.

Section 8 of 8

Hour-by-hour: the first 24 hours

Hours 1-2

  • Create the Next.js repo, deploy an empty page to Vercel.
  • Provision Supabase and create the tutors, students, reports and credits tables.

Output: Live URL and a schema ready for writes.

Hours 3-5

  • Seed the 12 levelled passages as JSON.
  • Build the intake form with passage picker and a file-upload field.

Output: An intake screen that accepts a student and a script.

Hours 6-8

  • Wire MediaRecorder for in-browser phone recording.
  • Push audio into Supabase Storage and store the path on the report row.

Output: A 60-second clip recorded on a phone and retrievable.

Hours 9-12

  • Integrate the managed speech API with word-level timestamps and cache the response.
  • Write the token-diff alignment into correct, substituted and skipped lists.

Output: A printed alignment result for one real clip.

Hours 13-18

  • Build the review-and-correct table with per-word overrides.
  • Build the first PDF template with a headline band and the three word lists.

Output: One downloadable branded PDF from a real recording.

Hours 19-24

  • Wire Stripe Checkout for the $19/$79/$199 packs and connect the account to TrustMRR.
  • Run an end-to-end test in test mode, then flip to live keys.

Output: A live payment path with zero customers, ready for launch.

Complete report

Full 30-day launch report

Read the full generated report

SpeakGrade Report Cards — 30-Day Launch Report

Concept version 1. JEV weighted score 6.62/10, band B, confidence 0.73. Highest criteria: speed to first revenue (8.23) and solo build feasibility (7.49). Lowest: monetization potential (5.34) and operational simplicity (5.14). This plan is shaped by all four.

1. Product opportunity and positioning

The buyer is the owner-tutor of a small English tuition business teaching children aged 6-14 in India, Pakistan, Vietnam, Indonesia or the Philippines. They invoice parents directly, usually on WhatsApp, at roughly $40-100 per child per month. There is no procurement cycle, no committee and no brand requirement to reach them — which is exactly why a solo founder can sell to them in week two.

The trigger is renewal. Every month or term end, the tutor must answer "what exactly did my child learn for that money?" The honest current answer is a voice note and "good progress!". Parents compare fees across tutors, then haggle or leave quietly. Producing individual progress notes for 25-40 students eats a weekend and still reads as opinion, so most tutors skip it and lose renewals without ever seeing why.

SpeakGrade's output is the artifact that renewal decision hangs on: a one-page PDF in the tutor's own name and logo, built from two 60-second reading recordings taken a term apart. It shows the words read correctly, substituted and skipped, four fluency metrics, a "words fixed since last report" list quoting the exact words, three drills and a next-term goal. The document is the tutor's, not ours — that is the point.

The original product sold a child-facing AI companion with a curriculum to a cold parent audience. SpeakGrade flips the buyer to the micro-business owner and shrinks the promise to one document. There is no AI character, no lesson content, no child login, and no rehearsal for a specific school assignment.

Positioning: the report card a tutor hands a parent at renewal, generated in three minutes from a phone recording.

2. Smallest sellable MVP

One flow, one output. A tutor opens a mobile web page, adds a student, picks one of 12 built-in levelled passages or pastes their own, records 60 seconds on the phone, and reviews the flagged words. The app transcribes with word-level timestamps, aligns the transcript to the script, computes words-per-minute, pause ratio, filler count and longest fluent stretch, diffs those against that student's previous report, and renders a one-page branded PDF that is emailed and shareable by WhatsApp link.

Sold as a $19 Trial Pack of five reports. Nothing else ships: no child-facing practice, no curriculum, no exam prediction, no native app, no free tier.

3. Technical implementation plan

A single Next.js app on Vercel with Supabase for Postgres, Storage and auth. The speech API is called once per recording and the timestamped JSON is cached, so re-rendering a report never pays twice. Alignment is a word-level diff in TypeScript. A review table lets the tutor override any flagged word before the PDF renders. Credits decrement on generation. Stripe Checkout issues packs and the account is connected to TrustMRR from day one, before any customer exists.

The hardest technical risk is not the build — it is speech-to-text mislabelling non-native child pronunciation and printing a wrong word into a document a parent reads. That is why the review step is load-bearing, and why day 20 exists to push median review time under two minutes.

4. Pricing and monetization

Three one-time credit packs: Trial at $19 for 5 reports, Term at $79 for 30, Centre at $199 for 100. No subscription. The Term Pack is the assumed unit. A $79 pack protects a $40-100 monthly tutoring fee, so no ROI argument is needed at the point of sale.

The honest weakness is repeatability: one-time packs can be a spike rather than a business. The mitigation in the plan is to log every buyer's next parent-meeting date and nudge them 25 days later, plus a referral bonus of 10 reports per introduced tutor.

5. First-customer acquisition

Direct Instagram DM is the only channel. Build a 150-row list in three days from tutor hashtags and Google Maps listings across Bangalore, Pune, Hyderabad, Lahore, Ho Chi Minh City and Jakarta. Before sending anything, generate a sample report card carrying that tutor's own name and logo. The DM asks permission to send the sample, not for money. The offer is five free reports for current students in exchange for one 60-second clip, delivered within 24 hours. The $79 Term Pack is asked for only after the tutor has handed the first PDF to a parent. Follow up on day 3 and day 7, once each.

6. 30-day execution roadmap

Days 1-5: build and ship the MVP, with day 5 as the hard deadline and live payment live from day 1. Days 6-8: proof assets and a 150-row list. Days 9-14: 60 personalised DMs, free packs delivered, first paid Term Pack targeted at day 10 by the plan, realistically days 10-14 depending on reply speed. Days 15-21: scale to 100 DMs, open Centre Pack conversations, launch the referral loop, and cut median review time under two minutes. Days 22-28: proof page with a paying tutor's permission, a fresh 40 DMs, term-end urgency messaging and landing-page fixes. Days 29-30: close all warm threads, reconcile Stripe against TrustMRR, and write the next-30-day plan. Full day-by-day tasks sit in the structured roadmap.

7. Revenue and profit scenarios

Three projections, all hypotheses. Conservative: 4 packs at $79, $316 gross, $16 in processing fees, $22 operating cost. Base: 16 packs, $1,264 gross, $63 fees, $55 operating. Upside: 30 packs, $2,370 gross, one $79 refund, $130 fees, $120 of paid Maps-keyword spend and $95 operating.

Contest ranking uses hackathon profit: gross revenue minus refunds, chargebacks, payment fees and paid marketing — hosting and speech-API delivery costs are excluded. Business profit subtracts operating cost on top. At base, hackathon profit is roughly $1,201 and business profit roughly $1,146. These are projections from stated assumptions, not verified revenue.

8. Main risks and validation experiments

Five risks carry kill criteria. False phonetic flags in front of a parent — tested by running 10 real child clips and timing corrections. DM conversion below assumption — tested by splitting 30 personalised against 30 plain-text sends. Tutors accepting free packs but never sending a clip — tested by offering 10 and counting clips within 48 hours. Read-aloud being called out as not real speaking — tested by showing five tutors the PDF and asking if they would hand it over as written. One-time packs not recurring — tested by logging every buyer's next parent-meeting date.

9. First 24-hour action plan

Six blocks. Hours 1-2: repo, deploy, Supabase schema. Hours 3-5: 12 passages and the intake form. Hours 6-8: phone recording into Storage. Hours 9-12: speech API plus alignment diff on one real clip. Hours 13-18: review table and the first branded PDF. Hours 19-24: Stripe Checkout and the TrustMRR connection, flipped to live with zero customers so day 9's first DM meets a working buy button.

This brief was generated from public data about established startups, then narrowed and scored for a solo 30-day launch. Revenue and profit figures are scenario models, not verified results, guarantees or evidence of existing customers. Validate demand yourself before you spend a month on it — and check the contest rules for what counts toward your score.

Related ideas

All ideas →