Contributor dashboardWelcome page, contribution summary and recording entry point.
Assamese Speech Technology Initiative
Give every Assamese voice a place in AI.
অসমীয়া ভাষাৰ কণ্ঠ
AxomVoice AI is building the speech-data foundation and technology needed for Assamese speech recognition, transcription and voice-enabled AI applications.
The working prototype
A platform to collect and validate Assamese speech
The current portal supports contributors in recording speech and reviewers in checking audio and transcripts. These tools form the starting point for building a quality Assamese speech dataset.
Guided speech recordingContributors read Assamese prompts or speak naturally.
Recording validationReviewers listen to audio and verify the transcript before acceptance.
From speech data to useful AI
Three steps towards Assamese voice technology
The long-term project combines structured data collection, model development and practical applications.
01
Collect speech
Gather Assamese recordings from contributors and prepare transcripts through a structured review process.
02
Train and evaluate
Use the curated dataset to train or fine-tune speech recognition models and evaluate them using WER and CER.
03
Enable applications
Develop speech-to-text for live speech, recordings and uploaded audio, with APIs for chatbot and AI integration.
Contribute your voice to Assamese AI.
Open the AxomVoice AI portal to record Assamese speech, contribute to the growing language resource and support the development of Assamese speech technology.