1
1+
2
2+
3
3+
4
Where most serious learners are stuck  ·  Where this gets you

Authentic texts.
Rated comprehension practice.
No ceiling.

Dingolab turns any real-world article, broadcast, or video into a GLOSS-style comprehension exercise — questions in English, passage in your target language, every answer tied to the exact line that proves it. Modeled on the DLPT. Built for learners who've outgrown everything else.

Join the waitlist See what's coming

Pre-alpha · Russian · More languages coming · ILR / CEFR / ACTFL

Why it's different
📐
Measurement, not acquisition

LingQ helps you read. Duolingo builds habits. Dingolab tells you what you actually understand — and where you break down — using the same ILR framework as the DLPT.

📰
Any authentic source

Paste text, drop a URL, or upload audio or video. The engine generates exercises from real content you choose — not simplified, scripted material written for learners.

🔗
Every answer has a source

After you submit, every option — right or wrong — is tied to the exact sentence that makes it correct or incorrect. No vague explanations. No guessing why you got it wrong.

Roadmap
01
Core Engine & Pre-Alpha
Reading comprehension · Russian · Engine validation · Closed beta
In Development Free · invite only

Building and validating the engine against real ILR-rated material before anyone outside the team touches it. Russian across five genres — journalistic, official, editorial, literary, technical — is the validation target. The engine ships to a small closed beta only after it passes that bar.

Live now
Paste any Russian text → rated comprehension exercise
Multiple-choice questions in English, passage in Russian
Estimated ILR level per question
Line-tied feedback on every answer option
ILR / CEFR / ACTFL display — your choice
Progress tracking across sessions
Level correction — flag mis-rated questions
Language status badges — Verified / Testing / Coming Soon
Feedback channel in the app
Privacy policy and terms of service
Error monitoring and analytics
Early access pricing page
Before closed beta opens
Item quality validation (manual gate)
First 20–30 waitlist invites sent
Now → Summer 2026 · First 20–30 waitlist spots open once Russian is validated
Join the waitlist →
02
Closed Beta · Intelligence & Input
URLs · Audio · Spaced repetition · First paid tier
Planned Pro · $12/mo

Expand what you can bring in and make the practice loop smarter. Paste a URL or upload audio and the transcript becomes your passage. Spaced repetition returns you to exercises where you had gaps. Export the vocabulary that cost you points straight to Anki. The Pro tier opens here.

What's included
URL import — paste any article link
Audio and video upload with auto-transcription
Spaced repetition on comprehension exercises
Genre tagging — journalistic, official, editorial, literary, technical
Accuracy breakdown by genre and ILR level
Vocabulary export to Anki from missed questions
Text-to-speech audio on reading passages
"Too easy / too hard" calibration signal
Basic weakness summary — where you break down by type
Late 2026 · Pro tier ($12/mo) and first Stripe billing launch here
Join the waitlist →
03
Soft Public Launch
Listening · Ukrainian · Adaptive onboarding · Content library
Planned Premium · $24/mo

The listening modality arrives — synchronized playback tied to the exact segment a question references. Ukrainian launches fully validated. New users get a short adaptive assessment that estimates their starting level before their first exercise. A library of open-licensed VOA and Wikisource content means no import required to get started. Public launch.

What's included
Listening exercises with segment-by-segment playback
Ukrainian — full reading and listening, Verified status
Adaptive onboarding — estimated ILR level on signup
Open-licensed content library — no import needed to start
Deep weakness fingerprinting — linguistic features, not just types
Mobile-optimized review mode
Validity evidence — real DLPT/OPI outcome data from users
Early 2027 · Premium tier ($24/mo) and full public access launch here
Join the waitlist →
04
Full Commercial Launch
Exam Prep · More languages · Account management
Planned $20 / exam cycle

A structured six-week Exam Prep program tied to your actual test date. The engine builds a practice schedule around your weak areas, surfaces the question types most likely to cost you points, and projects a score range from your accuracy trajectory. Buy it once per test cycle — no subscription required. Spanish, French, Arabic, and Mandarin verification also land here.

What's included
Exam Prep mode — $20/cycle, any tier, no subscription
6-week schedule built around your weakest question types
Weekly diagnostic emails — what to focus on next
Predicted score range from accuracy trajectory
Top-up credits — buy extra exercises without upgrading
Spanish, French, Arabic, Mandarin — Verified status
Full billing dashboard — upgrade, downgrade, cancel in one click
2027–2028 · Contingent on validated outcome evidence from Phase 3
Early Access

Early Access — Limited to 200 seats

Lock in your price before public launch. Grandfathered forever.

Loading seats…
FOUNDER · FIRST 100 SEATS
Founder
$5/mo, billed $60/year
Save $12/year vs monthly
  • 25 credits/month
  • All models — Haiku (1 cr) · Sonnet (2 cr) · Opus (3 cr)
  • Credits reset on the 1st each month
Grandfathered forever · Never pay more
EARLY ACCESS · SEATS 101–200
Early Access
$7/mo, billed $84/year
Save $24/year vs monthly
  • 35 credits/month
  • All models — Haiku (1 cr) · Sonnet (2 cr) · Opus (3 cr)
  • Credits reset on the 1st each month
Grandfathered forever · Still below launch pricing
At public launch, Pro will be $12/month ($108/year) with 60 credits. Early access members keep their price permanently.
Price guaranteed as long as your subscription stays active. Cancel anytime. Resubscription at current pricing.
Pricing
ALWAYS FREE
Reader
$0 / forever
No credit card. No time limit.
  • 10 exercises per month
  • Full line-tied feedback
  • ILR / CEFR / ACTFL display
  • Library saves for 7 days
  • 30 days progress history
MOST POPULAR
Pro
$12 / month
$9/mo billed annually — save $36/yr.
  • 60 exercises / month
  • Up to 20 credits roll over
  • Auto model routing (all ILR levels)
  • Full library & progress history
  • URL import
  • Vocabulary export to Anki
POWER USERS
Premium
$24 / month
$18/mo billed annually — save $72/yr.
  • 150 exercises / month
  • Up to 50 credits roll over
  • Full model control (force Opus)
  • Weakness fingerprinting
  • Exam Prep included
  • Early access to new features
ADD-ON · ANY TIER
Exam Prep
$20 / cycle
One-time. 30 Opus credits, valid 8 weeks.
  • 6-week structured schedule
  • Targeted weak-spot practice
  • Predicted score range
  • Opus model on all prep exercises
  • No recurring commitment
Our commitments

What we will always do.

Honest level estimates. ILR levels are labeled "estimated" everywhere in the app. They are calibrated against real user outcomes over time and will never be claimed as official or certified. When our estimates are wrong, we want to know — every correction you submit improves the engine for everyone.

No pressure to pay. The free tier gives you the core product — not a crippled version of it. Paid tiers add features that compound over time, not restrictions designed to frustrate you into upgrading. You will never be asked for a credit card to try the basic experience.

Privacy-first for serious users. We know many learners in this community work with sensitive material. Dingolab is designed for open, publicly available sources only — and we do not store or redistribute the content you bring in. Generated exercises (questions, feedback, your progress) are yours. Your source material passes through transiently.

No dark patterns. Cancel in one click — no guilt screens, no "are you sure," no hidden steps. If you cancel, your access runs through the end of the period you paid for, then drops to the free tier. Your library and progress data are never deleted on downgrade.

FAQ
Why not just ask ChatGPT or Claude to make me questions?

You can — and for casual practice, go for it. Dingolab exists because doing it well, repeatably, is a different problem than typing "give me questions about this text." A raw chatbot has no fixed standard, no memory of how you did last week, and no way to stop itself from inventing a justification that isn't in the text. Dingolab is built specifically to fix those three things.

What does Dingolab actually do that a chatbot doesn't?
  • Rates against a real scale. Every question is tagged with an estimated ILR level using the same reading-proficiency rubric the DLPT and GLOSS are built on — not a vague "easy/hard." You can see the exact rubric the engine rates against.
  • Verifies its own answers. Each option is tied to the exact sentence in the passage that proves it, and the engine checks that quote is really there — fabricated citations are caught and thrown out before you see them. A chatbot will confidently make one up.
  • Measures you over time. It tracks your accuracy by question type and by ILR level across every session, so you can see where you break down. A chat thread forgets everything the moment you close it.
  • Designs distractors deliberately. Every wrong answer targets a specific, named misreading, and options are matched in length and form so the answer isn't a giveaway. Ad-hoc chatbot questions are often guessable without reading the text.
  • Improves from real feedback. Questions rated well by users and reviewers are fed back in as the quality bar, so the engine gets better at your kind of material. A fresh chat starts from zero every time.
Is this a tutor or a test?

A test — by design. Dingolab measures comprehension against a rated scale; it doesn't teach you the language or read for you. The passage-anchored feedback shows you why each answer is right or wrong after you submit, but the point is to assess what you actually understand, GLOSS/DLPT-style.

Are the ILR levels official?

No. They are labeled "estimated" everywhere on purpose — the engine estimates difficulty against the ILR rubric, it does not certify it. Estimates are calibrated over time against real user outcomes and reviewer corrections, and will never be claimed as an official rating.

What languages and material does it work with?

Right now: reading comprehension in Russian, from any authentic text you paste. Ukrainian, URL/audio/video import, and listening are on the roadmap above. Dingolab is built for open, publicly available sources — your source text passes through transiently and isn't stored or redistributed.

Get early access.

Be among the first users when the closed beta opens. You'll shape what gets built first and hear about each phase before it goes public.

No spam. Unsubscribe anytime. We'll email when early access opens.
Privacy Policy · Terms of Service