Avenzoar
Play · Earn · Validate

We collect
data by
making games.

A client comes to us with a data need - the weather, a dialect, dog breeds, anything. We turn it into a game (or build a new one), real people play it, and their answers become clean, annotated data.

1
1
1
1
One round paid
+120 coins
Correct! +50
7-day streak
It looks like trivia
it's really data collection
Players earn real cash
for sharing what they know
Validated by the crowd
many players cross-check each answer
Delivered to clients
clean, tagged, ready to use
How it works

Play. Validate. Earn.
That's the whole loop.

It's the tagline and the business model in one: players have fun and get paid; the crowd's agreement turns play into trustworthy data.

01

Play.

Pick a game - anything from diacritics to dialects to dog breeds. Sessions are 30 seconds to a few minutes. It just feels like trivia.

  • No surveys
  • No queues
  • Any topic, any game
02

Validate.

Behind the scenes, the same item is shown to several players across different games. Where the crowd agrees, the label is trusted. Where it splits, it goes to a tie-breaker.

  • Many eyes per item
  • Agreement = quality
  • Gold checks catch bad actors
03

Earn.

Correct, agreed-upon answers mint coins. Convert them to local currency, mobile money, or gift cards.

  • Posted rates - no surprises
  • Bank · mobile money · gift cards
  • Cash out for real
Two sides, one platform

Players have fun and earn. Clients get data.

The arcade

Every game is
secretly collecting data.

These are formats we've already built. Drop your own data into one - or we design a brand-new game around your task. Hover any game to see what it collects.

Diacritization
Tashkeel
Tashkeel
Diacritization
flip
1
Diacritization

Tashkeel

Read the sentence, tap the right diacritic on each letter.

Written Arabic usually drops short vowels - so the same letters can be many different words. Restoring diacritics (tashkeel) is one of the hardest, highest-value problems in Arabic NLP.

text · MSA + dialects
OCR / handwriting validation
OCR Check
OCR Check
OCR / handwriting validation
flip
1
OCR / handwriting validation

OCR Check

See the original scan and the machine reading. Match or wrong?

Arabic OCR - especially handwriting and historical manuscripts - is notoriously unreliable. Human verification turns raw OCR into gold transcription data.

vision + text
Response quality rating
Answer Scale
Answer Scale
Response quality rating
flip
1
Response quality rating

Answer Scale

Weigh a question and its answer: excellent, partial, or wrong.

Rating answer quality is the backbone of RLHF and model evaluation. A balanced crowd of native speakers produces the preference data that aligns models.

preference · RLHF
Dialect identification
Dialect Compass
Dialect Compass
Dialect identification
flip
1
Dialect identification

Dialect Compass

Read the text, point the compass at its dialect.

Arabic is dozens of distinct spoken dialects, not one language. Labeled dialect data is scarce - and essential for models that understand how people actually talk.

Gulf · Egyptian · Levantine · Maghrebi
New game formats ship regularly.
Any data a client needs - if we can turn it into a 30-second game, we can collect it.

Who are you?