Idea AI & ML

Assamese Voice AI & Speech-to-Text Engine

Fine-tuning OpenAI Whisper on crowdsourced Assamese voice recordings to create an open-source speech recognition model.

The Problem

Voice interfaces (like Siri or Google Assistant) fail to understand native Assamese accents and regional dialects.

The Idea

Fine-tune OpenAI Whisper or Wav2Vec2 on crowdsourced Assamese voice data to create a lightweight, open-source Speech-to-Text model that runs locally on smartphones.

Status & Roadmap

💡 Community Idea — Coming Soon! This is an open community proposal. Maintainers and audio contributors are currently onboarding. We will update real code repositories, benchmark datasets, and live demos once ready and available.

Explore Other Projects

Check out our live tools and upcoming community proposals.

Coming Soon

Axomiya AI Tutor — LLM Assistant for Learners

An interactive LLM learning assistant that translates programming concepts and error logs into be...

Live

Krittimat — AI Tools & Prompts Hub

Know all the most famous AI tools for everything — from Image Generation to Voice Cloning, Video ...

Coming Soon

Open Assam Data Hub — FOSS Open Datasets

Aggregating, cleaning, and publishing open-license Assamese speech, text, dictionary terms, and r...

Beta

WhatAfter12 - Career Clarity Guidance

Know all higher education path, jobs & earning opportunities available after 12 in the age of AI.

← Back to All Projects