High-performance Coqui TTS API server with a hybrid "Hot/Cold" worker architecture
-
Updated
May 15, 2026 - Python
High-performance Coqui TTS API server with a hybrid "Hot/Cold" worker architecture
High-performance Whisper STT API server with a hybrid "Hot/Cold" worker architecture.
Reproducible TTS/STT benchmarks for the Uttera voice stack — corpora, harness, raw results, and decision guidance.
High-throughput STT server based on vLLM continuous batching. Whisper-large-v3-turbo and future Transformer STT models. RTF 400x+ on a single RTX 5090.
High-throughput TTS server based on vLLM continuous batching. VoxCPM2 and future Transformer TTS models. Optimized for cloud deployment and multi-tenant serving.
To associate your repository with the uttera topic, visit your repo's landing page and select "manage topics."