Features

One upload becomes a complete, phonics-correct listening lesson.

Every capability below is part of the same pipeline — you don't stitch five tools together. Bring a coursebook; leave with drills your students can practise anywhere.

① Ingest anything

The books you already own — in any format.

Every common file type

PDF (native and scanned), PowerPoint (PPTX), Word (DOCX), and image uploads — PNG, JPG, WebP. Up to 25 MB per file.

A photo of the page works

Snap a textbook page on your phone and upload it. No scanner, no clean PDF, no re-typing required.

Structure preserved

Headings become sections, slides stay slides, and column-aware tables are recovered as structured vocabulary rows.
② Extract & classify

It reads the page — then understands it.

AI-powered OCR

Scanned or image-only pages are read by an AI vision model with automatic fallback, so even a photographed textbook becomes editable text.

Text-trust re-OCR

Pages with garbled CID fonts or mojibake are detected automatically and re-read — no more silent “gibberish in, gibberish out”.

Pedagogical classification

Each item is typed: title, vocabulary word, example sentence, comprehension question, answer. The audio pipeline then knows how to drill each one.

Confidence & review queue

Low-confidence items are flagged into a keyboard-fast queue (a approve, e edit, x reject) — so nothing wrong reaches students.

No silent failures

A lesson that recovered no real content is marked “needs review”, never a false “100% done”. One click re-extracts it.

Quality score per lesson

Every extraction gets a quality score and flags, recorded per run — so you always know how well a book was read.
③ Drill — the moat

Phonics-correct audio, not generic narration.

This is what sets Languages Lab apart: audio engineered for young learners to hear the sounds, not just a smooth voice reading a paragraph.

IPA / SSML precision

A precision speech engine driven by strict SSML with IPA phoneme tags and timed breaks — each syllable and sound rendered clearly for phonics practice.

Multi-engine routing

Drills use the strict phonics voice; storytelling and dialogue use a natural, expressive voice. The right engine for each pedagogical job.

Real drill modes

Word-repeat, sentence-repeat, AI-chunked repeat-after-me, and backchaining — the patterns teachers actually use in class.

Karaoke highlighting

Words light up in time with the audio, using real word-boundary timing — so learners follow along and read as they listen.

Learner-tunable

Adjustable playback speed and repeat-pause length, remembered per learner — slow it right down, then speed up.

Vocabulary drilling

Vocabulary rows become full word-drills; example sentences become sentence practice — automatically, from the table on the page.
④ Deliver anywhere

Online, or completely offline.

In-app student viewer

A clean web viewer with per-item players, drills, images, and karaoke — the same look and feel as the studio, light or dark.

Self-contained offline bundle

Export one HTML file with every image and audio clip embedded. Students open it from a USB stick or shared folder — no login, no app, no internet. Ideal for low-connectivity classrooms.
⑤ Run it like a business

Built as real multi-tenant SaaS.

Multi-tenant isolation

Every school's data is isolated at the database level (row-level security), proven with automated tests.

Per-seat billing

Self-serve plans and subscriptions via Stripe, with server-side quota enforcement and a usage dashboard.

Usage & cost ledger

Every AI call — OCR, classification, TTS — is metered per tenant, so spend is transparent and attributable.

Privacy-aware

GDPR/COPPA-conscious design for a young-learner audience: data export, erasure, and retention controls.

See it with your own book.

Upload one page and hear phonics-correct drills come back. Free to try, no card.