{"name":"ming-voice","repo":"JMS1717/ming-voice","url":"https://midorreal.com/project/jms1717-ming-voice","repository":"https://github.com/JMS1717/ming-voice","description":"Low-latency local speech-to-speech voice assistant with Ming-3.0-audio-tiny on AMD and Nvidia GPUs. Easy setup on Windows and Linux","facts":{"license":{"value":"MIT","source":"https://github.com/JMS1717/ming-voice","checked":"22 Nov 2026"},"language":{"value":"Python","source":"https://github.com/JMS1717/ming-voice","checked":"22 Nov 2026"},"last_release":{"value":"v1.9.0, 18 Nov 2026","source":"https://github.com/JMS1717/ming-voice/releases","checked":"22 Nov 2026"},"releases":{"value":"10","source":"https://github.com/JMS1717/ming-voice/releases","checked":"19 Nov 2026"},"contributors":{"value":"1","source":"https://github.com/JMS1717/ming-voice/graphs/contributors","checked":"19 Nov 2026"},"archived":{"value":"no","source":"https://github.com/JMS1717/ming-voice","checked":"22 Nov 2026"}},"stars":451.0,"writeup":{"what_it_is":"Ming Voice is a local, real-time voice assistant built on the Ming-3.0-audio-tiny speech-to-speech model. It runs on AMD Radeon GPUs (NVIDIA is listed as experimental) on Linux or Windows via WSL2. It listens, detects when the user has finished speaking, and replies out loud, with tools for weather, web search, timers, lists, and Home Assistant control.","audience":"Home users wanting a self-hosted voice assistant on their own GPU","claims":[{"kind":"specific","claim":"Answers start about a quarter of a second after the user stops talking","excerpt":"usually starting **about a quarter of a second** after you stop talking","status":"not_checked"},{"kind":"specific","claim":"Runs fully local with no audio leaving the machine","excerpt":"**No audio leaves your machine.**","status":"not_checked"},{"kind":"specific","claim":"Reaches about 115 tokens/s for the thinker model on a Radeon RX 7900 XTX","excerpt":"Thinker (audio LLM, 9.3B MoE / 2.7B active) | ~115 tokens/s","status":"not_checked"},{"kind":"specific","claim":"Needs about 20.5 GB VRAM in bf16, about 12.5 GB with int8 experts","excerpt":"~20.5 GB in bf16, ~12.5 GB with int8 experts","status":"not_checked"}],"alternatives":[],"written_by":"AI, from the project's README","date":"2026-11-17"},"why_now":null,"owner_supplied":null,"score":{"verdict":"sleeper","hype":27,"reality":75,"gap":-48,"momentum":null,"confidence":60,"version":"score-2.0","rule":"verdict-2.0","calculated_at":"2026-11-22T08:00:00+00:00","verdict_source":"rule"},"early_read":null,"changes":[{"date":"2026-11-18T16:38:59+00:00","text":"Latest release v1.9.0","source":"https://github.com/JMS1717/ming-voice/releases"},{"date":"2026-10-21T11:37:10+00:00","text":"First release","source":"https://github.com/JMS1717/ming-voice/releases"},{"date":"2026-10-21T01:08:35+00:00","text":"Repository created","source":"https://github.com/JMS1717/ming-voice"}],"cite":"ming-voice is a GitHub project at https://github.com/JMS1717/ming-voice written mainly in Python, described there as \"Low-latency local speech-to-speech voice assistant with Ming-3.0-audio-tiny on AMD and Nvidia GPUs. Easy setup on Windows and Linux\".","attribution":"Data from Mid or Real, https://midorreal.com/project/jms1717-ming-voice, checked 22 November 2026. Attribution with a link is required.","still_real":null}