SAASPOCALYPSEverdict #GLAASTER-C32F
scanned 2026.07.11 · 12:07
subject of investigation

glaaster.com

AI reading aid for DYS learners
verdictSOFT
wedge score
73
/100
wedge thesis

the door is distribution and pricing: they rely on institutional channels and a freemium demo, so reach and upsell are the weak points indie teams can attack.

wide-open walls — wedgeable·ship in 6 weeks·run for $1.00 + usage
the doornetwork
wedge

where the walls are.

methodology →
the door

no network effect to overcome — users don't compound users.

watch out

the technical wall is real — research-grade engineering, not a weekend.

capital
3.0/10
investment the incumbent had to make
why this scoremedium confidenceLow capital requirements and no proprietary hardware or inventory, but institutional sales and in-region hosting add...

Low capital requirements and no proprietary hardware or inventory, but institutional sales and in-region hosting add moderate costs.

  • Product uses standard cloud hosting and open-source OCR/TTS; estimated competing cost ~ $1 + usage.
  • Harder costs are institutional sales and region-specific hosting (French OVH) for privacy compliance.
  • No mention of proprietary infrastructure, large compliance teams, or inventory.
technical
4.0/10
depth of the underlying engineering
why this scoremedium confidenceCore features are composable from off-the-shelf OCR, TTS, and LLMs, so engineering depth is limited but integrations...

Core features are composable from off-the-shelf OCR, TTS, and LLMs, so engineering depth is limited but integrations and privacy choices add some complexity.

  • Core features listed: OCR, TTS, document upload, AI assistant — all solvable with existing libraries/APIs (Tesseract, Web Speech API, LLM APIs).
  • Notes call out medium effort for multi-language TTS and privacy/hosting requirements (OVH, non-training guarantees).
  • Detected stack: Next.js, Supabase, Tesseract — standard web stack.
networkdoor
1.0/10
users compound users
why this scorehigh confidenceNo evidence of marketplaces, user-generated content, or multi-sided liquidity; distribution is institution-focused...

No evidence of marketplaces, user-generated content, or multi-sided liquidity; distribution is institution-focused not network-driven.

  • Report explicitly notes moat is partnerships with schools and research legitimacy, not network effects.
  • No mention of marketplaces, social graphs, UGC, or partner/app ecosystem.
  • Distribution signals (knowledge_graph_present etc.) are null.
switching
2.0/10
stickiness of customer data + workflow
why this scoremedium confidenceLow data lock-in and standard document formats mean users can migrate; institutional procurement might add friction...

Low data lock-in and standard document formats mean users can migrate; institutional procurement might add friction but not product-level switching costs.

  • Features include file uploads and OCR supporting standard documents which are exportable and composable.
  • No mention of proprietary file formats, deep integrations, or irreversible data state.
  • Institutional contracts could create administrative switching friction per 'institutional sales' note.
data
2.0/10
proprietary data accumulates over time
why this scoremedium confidenceNo proprietary training corpus or unique behavioral dataset; claim of not using data for model training reduces data...

No proprietary training corpus or unique behavioral dataset; claim of not using data for model training reduces data capture advantages.

  • Report states non-use of data for model training and emphasis on privacy (in-region hosting).
  • Core functionality relies on standard OCR/TTS/LLMs rather than proprietary models or accumulated labeled data.
  • No mention of large behavioral or fraud datasets unique to the product.
regulatory
3.0/10
real licenses, not SOC 2 theater
why this scoremedium confidenceOperating in education and handling student data introduces regulatory and compliance needs (privacy, hosting), but...

Operating in education and handling student data introduces regulatory and compliance needs (privacy, hosting), but no high-bar licenses like HIPAA/FINRA are cited.

  • Report highlights French OVH hosting and privacy commitments for institutional customers.
  • Targeting schools and clinical validation implies some compliance and approvals may be required.
  • No mention of regulated licenses (HIPAA, FINRA, money transmission) or heavy regulatory obligations.
take

the blunt take.

Glaaster is a research-backed, language-specific assistive reader with sensible product-market fit in schools and families, but the public acquisition funnel is thin and gated behind institution-focused messaging — distribution, not tech, looks vulnerable.

Core features (layout adaptation, TTS, OCR, document upload, an AI assistant) are composable from off-the-shelf OCR, TTS and LLMs; the hard moat is partnerships with schools and research legitimacy, which are durable but addressable by targeted grassroots distribution and bilingual UX.

cost

cost of competing.

what they charge
inferred freemium + paid upgrades
free tier + paid features
/ account/mo
homepage shows free tier with limited advanced features; paid pricing not public
annual:depends on plan
what running yours costs
01 · Vercel (hobby tier)$0.00
02 · Supabase (auth + Postgres free → Pro when scaling)$0.00
03 · OCR (Tesseract self-hosted or Google Vision at small scale)$0.00
04 · TTS (Resemble/Resend prototype / or browser Web Speech API)$0.00
05 · LLM API for assistant (light use)??? — scales with usage
06 · Domain$1.00
TOTAL / mo$1.00 + usage
▸ break-even:immediately — the free tier forces you to differentiate on UX and access, and paid upgrades would pay for a $1/mo indie build as soon as one user converts to paid
build

what you're up against.

2 days: MVP OCR + upload · 1 week: adaptive reader UI and presets · 1 week: TTS + simple prompt-based assistant · 2 weeks: quizzes/flashcards + export · 1.5 weeks: polish, onboarding and privacy copy
easy
medium
hard
nightmare
01
easy
File uploads and OCR
Standard infrastructure and libraries (Tesseract or Vision API); ensure basic sanitization and per-account storage.
02
easy
Adaptive style presets (fonts, spacing, colors)
Pure front-end work: CSS variables and a settings profile per user; quick wins for perceived value.
03
medium
TTS integration and language support
Browser TTS covers basic needs; higher-quality voices require paid APIs and handling multi-language audio files.
04
medium
Simple AI assistant for definitions and reformulations
Prompt engineering and safety filters are required; usage costs can grow with active sessions.
05
hard
Document privacy & hosting in-region
Glaaster highlights French OVH hosting and non-use for model training — matching that requires storage choices and clear privacy copy.
06
nightmare
Institutional sales and research partnerships
Winning school contracts and clinical validation is slow, bureaucratic, and often requires formal studies or endorsements.
stack

their position.

detected signals· measured
cmsWordPress
recommended stack · inferred
inferNext.js (static + serverless routes)inferSupabase (auth + Postgres) or Neon freeinferTesseract OCR (self-host) + fallback Vision APIinferBrowser Web Speech API + optional Resemble/ElevenLabs (paid)
rivals

who else has tried this.

option A
Self-hosted Tesseract + simple web UI
cheap OCR + custom reader controls give control and privacy for small classrooms.
option B
Use Chrome/Edge Reader + native TTS
lower-tech path: leverage browser accessibility features and deliver value without building AI features.
option C
BeeLine Reader / dyslexia fonts + audio notes
combine existing accessibility tools to mimic most adaptation features without a full platform.
compare

similar scans.

same shape - different moat
ready to wedge in?
Get the wedge plan. Ship a contender by Sunday.
▸ generated with love, by a heartless robotverdict v2.5 · saaspocalypse.dev