BanglaLLMBanglaLLM logo mark
Sign inJoin in

About BanglaLLM

Teaching AI the Bangla we really speak

Bangla is spoken differently in every part of Bangladesh, yet AI still understands little of it. BanglaLLM is a free, volunteer effort to change that, one short written answer and recording at a time.

Join in now See an example task
—volunteers
—contributions submitted
—verified examples in the dataset
—local dialects, one goal

Why we do this

Our language deserves to be understood as it is spoken

People across Bangladesh speak Standard Bangla and many regional dialects, each with its own words, sounds and sayings. Most of that richness is missing from the data AI learns from, so AI often misreads it or answers in a Bangla few of us speak at home.

Good data from real speakers is what closes that gap. So we ask people to write short answers in their own words and dialect, and we let people from the same region check them.

What this is, and is not

A dataset that improves existing AI models, not a new model

What we are doing

  • Collecting written answers and voice recordings in the Bangla people really speak, region by region.
  • Having people from the same region check the answers.
  • Building a dataset for research and for fine-tuning existing open weight models, which may be shared openly.

What we are not doing

  • We are not building a new AI model from scratch.
  • This is not a chatbot or an app to chat with.
  • We do not use your voice to imitate you.

How it works

From your keyboard to the dataset

  1. 1You write

    Each task is a question, a situation or a sentence to say your way. You type a short answer in Bangla script and read sentences aloud to record your voice, about a minute each.

  2. 2Automatic checks

    The system checks the basics: Bangla script, enough length, no copy of an earlier answer, and recordings that are not silent or too noisy.

  3. 3People review

    Reviewers who speak your dialect read the answer. Nothing is accepted without a person approving it.

  4. 4The dataset

    Accepted writing and recordings form a dataset that helps existing open weight models (for example Gemma and Qwen) understand and speak Bangladeshi Bangla.

Writing and voice

Writing

Written answers, conversations and sentences in your dialect. They teach AI to read and write our Bangla.

Voice

Recordings of you reading sentences aloud and speaking your answers, so AI can understand and speak our Bangla. You agree to recording in the consent form first.

Dialects across Bangladesh

Every region has a voice here

Choose the dialect you speak at home, or more than one. If yours is not listed, pick Standard Bangla or the closest one and write the way you really speak.

  • Standard
  • Dhaka & central
  • South & south-west
  • Chattogram & Cumilla
  • Sylhet
  • North-west (Rajshahi)
  • North (Rangpur)

Your trust, our responsibility

Your words stay yours

You choose to take part

You agree before you start, knowing what is collected and why.

Others see only your ID

Only you and the project team see your name; everyone else sees a random ID and your avatar.

Leave any time

Withdraw your consent and your recordings are deleted, your answers leave all later datasets and your name is deleted.

Free and voluntary

It costs nothing. Give as much or as little time as you like.

Getting started

From sign-in to your first answer in a few minutes

  1. 1Sign in with your email

    We send a one-time code to your email.

  2. 2Tell us your dialect and home

    Pick the dialect you speak at home and where you are from, so your words reach the right region.

  3. 3Read and agree

    The consent form says what is collected and why. You agree before anything is collected.

  4. 4Start your first task

    Write a short answer, or read a sentence aloud. About a minute, and you can stop whenever you like.

As you contribute, you may be invited to review other people's work from your own region. Reviewers see only an ID and an avatar, never a name. More questions? Read the FAQ.

Give a few minutes, give your language a future

Every answer you write helps AI learn the Bangla we really speak.

Start today
BanglaLLMBanglaLLM logo mark

My language, my AI

A data-collection platform. Your writing and voice recordings become a dataset that helps existing open weight models (for example Gemma and Qwen) understand and speak Bangladeshi Bangla. We are not building a new model from scratch.

The project

AboutHow it worksDialectsFAQ

Take part

Join inSign inSee an example task

Legal

Privacy PolicyTerms and Conditions
© 2026 BanglaLLM. A free, volunteer project.