Official documentation
fcitx5-vinput documentation: Push to talk, multi-engine ASR, AI rewriting
xifan2333, fcitx5-vinput documentation (GitHub Pages), 2026
Read the originalRead on October 25, 2026
- What it is
- The official documentation site for fcitx5-vinput, a voice input addon for Fcitx5 on Linux. Its home page lists built-in sherpa-onnx recognition, optional cloud recognition, and LLM-powered rewriting.
- How much weight it carries
- The maker's own documentation site, read on the date above. It is a feature overview, not a privacy policy or pricing page, so it says nothing about data retention or training.
Facts taken from this source
- The fcitx5-vinput documentation says its AI rewriting covers LLM correction, formatting, translation and instruction-based rewriting.
“AI rewriting Go beyond transcription with LLM-powered correction, formatting, translation, and instruction-based rewriting.” (AI rewriting)
- The fcitx5-vinput documentation says the built-in sherpa-onnx engine can recognize speech locally.
“Use built-in sherpa-onnx locally or switch to cloud ASR when you want different speed, quality, or cost.” (Multi-engine ASR)