openVoiceInput_linux README on GitHub

Facts taken from this source

• The project's README says its cleanup step does not call an LLM and adds no network request.

  “清理不调用 LLM，也不增加网络请求，不替换术语、数字或大小写” (Paragraph on the 'cloud recognition' page's cleanup modes)

• The project's README says that correcting a phrase, such as 'bench mark' to 'benchmark', creates a rule for that phrase only and not for the whole sentence.

  “例如把 bench mark 修成 benchmark 时，只形成这个短语的规则， 不会把整句一起记住。” (Paragraph on learning from exact edits)

• The project's README says it has no local speech recognition yet and does not register system shortcuts automatically.

  “本项目目前没有本地 ASR， 也不会自动注册系统级快捷键。” (Notice above the install section)

• The project's README says that being a light client does not mean the app is fully offline.

  “这是“轻量客户端”的含义，并不等于完全离线” (Section 'Small installer')