Shruti open-weights speech-to-text covers all 22 Indian languages offline in under 30MB
kingslyj · x · 2026-11-08
udgamlabs released Shruti, an open-weights (Apache-2.0) speech-to-text model covering all 22 languages of India's Eighth Schedule. Each language fits in a file under 30MB — about a fifth of IndicConformer's 137.7MB — and runs offline in the browser on a five-year-old phone. Built on Cactus Whistle, it targets developers who want to serve millions of Indian users without big-company dependence, cloud bills, or large GPUs.
More from Infra
- 10,000 Starships a Year: SpaceX 2030 Target vs All 2,533 US Launches Ever — LazyHomoSapiens · 2026-11-08
- Dev runs SmolLM-135M inference entirely in SystemVerilog via Verilator — cneuralnetwork · 2026-11-08
- How Robert Dennard's 1966 One-Transistor Idea Became the Foundation of Modern DRAM — cneuralnetwork · 2026-11-08
- Inkling 2's KV cache is 10x DeepSeek's, fueling doubts about RSI-driven architecture gains — bookwormengr · 2026-11-08
- Anthropic API users beware: cache 5-minute window can make bundled credits hilariously wasteful — teortaxesTex · 2026-11-08
- Harbor turns one: 200T tokens/week, 10M+ Terminal-Bench runs, 1B+ rollouts — BenBlaiszik · 2026-11-08