Welcome to 0N Labs, where we pioneer lightweight, on-device AI models for text, speech, and reasoning. Our mission is to shrink cutting‑edge algorithms into efficient, high‑quality systems that run on everyday hardware—no expensive GPUs required.
- Democratizing AI: Make advanced language and voice intelligence accessible on phones, wearables, IoT devices, and microcontrollers.
- Efficiency First: Design and optimize models to minimize memory, compute, and energy while preserving performance.
- Real‑World Impact: Deliver solutions that respect privacy, operate offline, and empower users in diverse contexts.
A neural text‑to‑speech engine with:
- 82M parameters and a 300 MB footprint
- Multi‑speaker, multilingual support
- Real‑time synthesis on CPU/NPU (RTF < 0.3)
- Advanced phonemization with IPA and stress markers
- Standardized tests for latency, memory, energy consumption
- Perceptual evaluations (MOS, intelligibility scores)
- Continuous integration with performance regression checks
We welcome contributions—from code optimizations to new model ideas and evaluation frameworks.
- Fork the repo & create a feature branch.
- Implement or tweak, then add tests or benchmarks.
- Submit a pull request with clear description and benchmarks.
- Careers & Internships: If you’re passionate about efficient AI, email your résumé and research CV to careers@0nlabs.ai.
- Partnerships & Investment: Interested in supporting or integrating our work? Contact partners@0nlabs.ai.
All code and models are released under the Apache 2.0 license. See LICENSE for details.
0N Labs — Making AI fit for the edge.