<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>The Repo Dude — Speech &amp; transcription</title><link>https://www.therepodude.com/updates.html</link><description>Dated editorial picks and documentation checks, not inferred releases.</description><atom:link href="https://www.therepodude.com/feeds/speech.xml" rel="self" type="application/rss+xml"/><item><guid isPermaLink="false">research-openai/whisper-2026-11-19</guid><title>Requirements checked: openai/whisper</title><link>https://www.therepodude.com/repos/openai/whisper.html</link><pubDate>Thu, 19 Nov 2026 00:00:00 GMT</pubDate><description>Python, PyTorch, model files and the FFmpeg command-line tool. The local package transcribes with downloaded weights; this is separate from a hosted transcription API.</description></item><item><guid isPermaLink="false">research-systran/faster-whisper-2026-11-19</guid><title>Requirements checked: systran/faster-whisper</title><link>https://www.therepodude.com/repos/systran/faster-whisper.html</link><pubDate>Thu, 19 Nov 2026 00:00:00 GMT</pubDate><description>Python and CTranslate2; PyAV bundles FFmpeg libraries. GPU acceleration needs compatible NVIDIA libraries. CPU and GPU inference run locally without a mandatory transcription API.</description></item><item><guid isPermaLink="false">research-ggml-org/whisper.cpp-2026-11-19</guid><title>Requirements checked: ggml-org/whisper.cpp</title><link>https://www.therepodude.com/repos/ggml-org/whisper.cpp.html</link><pubDate>Thu, 19 Nov 2026 00:00:00 GMT</pubDate><description>A compatible build or binary and model files; CPU-only inference is supported. Inference is local; no hosted transcription key is needed for the documented local workflow.</description></item><item><guid isPermaLink="false">research-m-bain/whisperx-2026-11-19</guid><title>Requirements checked: m-bain/whisperx</title><link>https://www.therepodude.com/repos/m-bain/whisperx.html</link><pubDate>Thu, 19 Nov 2026 00:00:00 GMT</pubDate><description>Python package, model files, alignment dependencies; CPU mode or compatible CUDA for GPU use. Local transcription is available. Speaker diarization requires a Hugging Face token and acceptance of the relevant model terms.</description></item></channel></rss>