About
Learn about the Enhancing Bangla Language in ICT (EBLICT) project through research and development.
Goal: To integrate the Bangla language into the global digital ecosystem through Artificial Intelligence and Natural Language Processing.
The 'Enhancing Bangla Language in ICT through Research and Development' (EBLICT) project, managed by the Bangladesh Computer Council, is a national initiative to create Artificial Intelligence, Data Science, and Bangla language software. Currently, this project serves as the hub for home-grown AI in Bangladesh. The main scope and achievements of the project are outlined below:
1. Key AI-based Applications
Kagoj.ai:
An integrated AI platform that assists in creating, processing, researching, and content creation of government and private documents. It includes Bangla OCR, STT, TTS, and spell checker.
Shothik:
An AI-driven spelling and grammar checker tool that follows Bangla Academy standards.
Borno OCR:
Converts images, PDFs, and handwritten documents into editable text. It supports computer-composed, typewriter, and letterpress documents.
Kotha:
A speech-to-text (STT) system that converts spoken Bangla into text.
Uccaron:
A text-to-speech (TTS) engine that reads written text aloud in a natural human voice.
Jiggasha.ai:
A Bangla virtual private assistant, connected with more than 50 government office services.
Note: Spell, OCR, and STT services are provided from this platform to a2i's D-nothi application.
2. Disability, Accessibility, and Inclusion
Special software has been developed for people with disabilities and underprivileged communities.
Bangla Screen Reader 'Alo':
Helps visually impaired people operate computers.
Sign Language Recognition System:
Converts sign language into written text.
Text-to-Sign Puppet:
Converts written text into animated sign language.
Braille Converter:
Converts Bangla text and mathematical symbols into Braille format.
3. Preservation of Ethnic Minority Languages
A platform has been created for the preservation of endangered ethnic minority languages.
multiling.cloud:
A digital repository of more than 40 indigenous languages of Bangladesh. It contains data of 14 endangered languages (e.g., Rangmitcha and Koda) and more than 12,000 minutes of audio.
4. Linguistic Resource and Research
Big data required for Bangla AI has been created.
Bangladesh National Corpus (BDNC):
A collection of more than 3 billion words and 20,000 digitized books, which will be used in LLM training.
BanBrain:
A language model capable of suggesting next words and identifying paraphrases.
Digital Lexicon:
The first technology-based Bangla dictionary.
Universal Keyboard – Uboard:
A custom layout creation platform supporting minority ethnic languages like Chakma, Marma, and Mro.
Purno and July Fonts:
Professional Unicode fonts for official use.
IPA Converter – Dhoni:
Converts Bangla text into International Phonetic Alphabet (IPA).
Unicode and CLDR Coordination:
Regular coordination is made with international standards to correctly display Bangla language and local culture such as Bangla date, time, measurements, and alphabet.
All these resources are currently available on bangla.gov.bd, various app stores (e.g., Google Playstore, Chrome Store), and the Huggingface account.
Through applicable services, home-grown apps can be used for free for general users in the B2C model, and there are opportunities for monetization through APIs in B2B channels for organizations.
Researchers, universities, and startup organizations can conduct research activities using GPUs through collaboration.
Note: The project tenure is July 2016 – June 2026 and the total allocation is 15896.69 lakh BDT (GOB allocation).