alumae/kaldi-gstreamer-server

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1093 stars · 338 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A real-time full-duplex speech recognition server built on the Kaldi toolkit and GStreamer framework, implemented in Python. It streams audio in over websockets and returns partial recognition hypotheses, with a scalable master-worker architecture.

Use cases

• stream live microphone audio and get real-time speech-to-text
• build a voice typing feature like Android's dictation
• transcribe long audio recordings with automatic segmentation at silences
• run a self-hosted speech recognition server with multiple concurrent sessions
• rescore recognition output with a large language model
• support many audio codecs for streaming ASR

When to choose

• you need self-hosted, real-time streaming speech recognition with partial results
• you want scalable recognition via independently deployable workers
• you already have Kaldi GMM or online DNN acoustic models

When to avoid

• you want a modern maintained ASR stack (e.g. Whisper or newer K2-based tools)
• you need Python 3.10+ or Tornado 5+ with post-processing, which has known issues
• you only need offline batch transcription without streaming

Facets

service · maturity maintenance

speech-recognition http-server websocket streaming speech-processing apis python self-hosted kaldi gstreamer asr voice-typing real-time-transcription audio-streaming natural-language-processing linux web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="alumae/kaldi-gstreamer-server")