Show HN: Nanosamur.ai – Ollama for Voice Models (STT) (github.com)

Nanosamur.ai has launched an open-source orchestration platform aimed at enhancing voice model capabilities for speech-to-text (STT) applications. This model-agnostic solution supports real-time, refined, and batch processing, allowing users to manage sensitive voice data within their own infrastructure. The platform features a robust suite of tools including a browser UI, a Windows-first Electron app, and a suite of APIs and SDKs, empowering developers to build and customize their own workflows around speech processing. Significantly, Nanosamur.ai addresses the increasing demand for privacy and control over voice data, making it appealing to organizations that prioritize data security while utilizing AI models. Key functionalities include speaker-aware transcription, workflow result integration, and multi-tenancy support for collaborative environments. The architecture supports various models like Faster-Whisper for real-time transcription and WhisperX for refined outputs, providing flexibility tailored to specific use cases. Additionally, the platform's use of Kafka for message broadcasting and PostgreSQL for transcript persistence underscores a modern approach to managing complex audio processing requirements while allowing for seamless integration into existing infrastructures.