Axomiya AI Foundry is a community-driven, open-source initiative by Axomiya IT Labs focused on building language resources, datasets, tools, and research infrastructure for Assamese AI.
We are working to make high-quality Assamese language resources more accessible to researchers, developers, educators, students, institutions, and open-source communities.
🌐 Visit Axomiya AI Foundry · 🏠 Axomiya IT Labs
Artificial intelligence is becoming part of everyday life—from search and education to accessibility, communication, research, and public services.
Modern foundation models such as OpenAI GPT, Google Gemini, Anthropic Claude, Meta Llama, Mistral, DeepSeek, Sarvam AI, and others are expanding multilingual AI capabilities.
These systems depend on high-quality language resources.
For Assamese, openly available and reusable AI-oriented resources remain comparatively limited.
Axomiya AI Foundry exists to help build that missing layer.
We focus on the language infrastructure beneath AI applications: structured datasets, speech resources, evaluation resources, documentation, tooling, and open research infrastructure.
We are not building another chatbot.
We are building resources that others can build with.
Build open, reliable, and reusable language infrastructure that helps Assamese thrive in the age of artificial intelligence.
We believe Assamese language technology can benefit from an open ecosystem where students, researchers, developers, institutions, and communities can contribute to and build upon shared resources.
Our work is focused on creating useful foundations that can support the next generation of Assamese AI applications and research.
An open structured Assamese lexical resource designed to support language research, search, retrieval, educational tools, RAG systems, and other language technologies.
Status: In development
A community-built Assamese speech resource containing recordings and verified transcriptions for speech technology research.
Status: In development
As the project develops, we plan to explore additional open resources including:
- OCR datasets
- Handwriting datasets
- Translation corpora
- Evaluation benchmarks
- Knowledge resources
- Multimodal datasets
- Speech resources
- Linguistic metadata
- Developer tooling
All resources will be published according to their individual availability, documentation, and licensing requirements.
Better language resources can support research and applications across:
- Large Language Models
- Retrieval-Augmented Generation
- Speech Recognition
- Text-to-Speech
- Machine Translation
- Search
- OCR
- Education
- Accessibility
- Digital Archives
- Language Learning
- Conversational Systems
Our goal is to make it easier for people to experiment, research, build, and innovate with Assamese.
We believe language infrastructure benefits from openness and collaboration.
Our work emphasizes:
- Open development
- Transparent dataset versions
- Community contribution
- Reproducible research
- Clear licensing
- Public documentation
- Dataset quality
- Long-term preservation
Everyone is welcome.
You can contribute as a:
- Native Assamese speaker
- Student
- Teacher
- Linguist
- Researcher
- Developer
- AI engineer
- Writer
- Translator
- Open-source contributor
- Community volunteer
You don't need to be an AI researcher to contribute.
A word, sentence, voice recording, correction, issue, pull request, or documentation improvement can all help strengthen the project.
Website, documentation, project information, research resources, and contribution guides.
Dataset engineering, schemas, validation tools, metadata, releases, and dataset resources.
Additional repositories may be created as the ecosystem grows.
Our infrastructure is designed around open technologies and interoperable formats.
Depending on the resource, this may include:
- JSON
- JSONL
- CSV
- SQLite
- SQLite FTS5
- WAV
- Python
- TypeScript
- Next.js
- GitHub
- Open-source AI tooling
Specific technologies and formats are documented within each project.
We aim to make our resources useful for academic and applied research.
Dataset releases will include, where applicable:
- Version information
- Dataset documentation
- Schema
- Licensing
- Statistics
- Validation information
- Changelog
- Citation information
Researchers should cite the specific dataset version used in their work.
Build Assamese language applications using open resources.
Potential applications include:
- Assamese search
- RAG systems
- Speech recognition
- Text-to-speech
- Translation
- Educational applications
- Accessibility tools
- Language learning
- Digital archives
- Conversational systems
- Language research
Axomiya AI Foundry is built in the open.
Explore our repositories, open an issue, suggest an improvement, contribute data, submit code, or participate in discussions.
Every contribution helps strengthen the Assamese open-source AI ecosystem.
Axomiya AI Foundry is an open-source initiative by Axomiya IT Labs, an Assam-based community focused on AI, open source, and accessible technology.
Axomiya IT Labs provides the broader community and open-source ecosystem in which projects such as Axomiya AI Foundry can be developed and shared.
🌐 Axomiya AI Foundry https://axomiya-ai-foundry.vercel.app/
🏠 Axomiya IT Labs https://axomiyaitlabs.vercel.app/
💻 GitHub Organization https://github.com/AxomiyaAIFoundry
📚 Documentation https://axomiya-ai-foundry.vercel.app/documentation
🔬 Research https://axomiya-ai-foundry.vercel.app/research
🤝 Contribute https://axomiya-ai-foundry.vercel.app/contribute
Axomiya AI Foundry Open infrastructure for Assamese AI.
An initiative by Axomiya IT Labs.