# fxy2311-youyou/expression-trainer

宇宙无敌表达训练系统 — 实时语音转文字 + 27000词情感词库 + AI表达分析报告。离线运行，对着它说话，它帮你看见自己的表达问题。

Repository: https://github.com/fxy2311-youyou/expression-trainer
Canonical: https://ross.abutalabs.com/products/expression-trainer
Language: JavaScript
License: MIT
License Family: permissive
Last push: 2026-07-18T10:03:37+00:00

## Health v2 (maintenance only)
Score: 55/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 93, release rhythm 35, longevity 4
- inputs: {"age_days": 58, "days_push": 46, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1105, forks 206 (observed 2026-08-28T04:03:34.840482+00:00)

## What it is
A local desktop app (Electron) that trains spoken expression skills by combining fully offline real-time speech recognition (Sherpa-ONNX) with a 27,000-word emotion lexicon to detect filler words, hesitation words, and vague terms. It provides live on-screen subtitles, real-time statistics, and AI-generated six-dimension analysis reports via configurable LLM backends (DeepSeek, OpenAI, Groq, Ollama).

## Use cases
- practice public speaking and reduce filler words like um and uh
- get real-time feedback on how clearly I express myself
- analyze my speech for vague wording and get precise replacement suggestions
- run an offline speech coach that transcribes my voice locally
- generate a detailed report on my speaking logic and word choice
- train interview or presentation speaking skills with AI feedback

## When to choose
- you want a privacy-friendly, offline speech trainer that processes audio locally
- you speak Chinese (or Chinese-English) and want lexicon-based filler/hesitation word detection
- you want live subtitles plus AI-generated expression analysis reports
- you prefer choosing your own LLM backend including free local Ollama

## When to avoid
- you need languages other than Chinese/English speech recognition
- you want a mobile or web-only solution without installing Electron and downloading ASR models
- you need fully free operation without any LLM backend setup
- you require enterprise-grade multi-user or team training features

## Facets
- artifact type: application
- maturity: active
- function: speech-recognition, nlp, llm-inference, chatbot, gui
- domain: developer-tools, education, desktop-applications
- platform: cross-platform, windows
- tags: speech-training, filler-word-detection, offline-asr, sherpa-onnx, public-speaking, emotion-lexicon, ai-feedback, natural-language-processing, electron, desktop, macos, linux

## Member repositories
- fxy2311-youyou/expression-trainer (main) score 55

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:34.840482+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:44:34.033353+00:00, confidence not recorded.
  - readme: https://github.com/fxy2311-youyou/expression-trainer (fetched 2026-08-28T04:03:34.840482+00:00, sha 39142b1979d3)
- Data as of 2026-08-30T08:39:29.467469+00:00.