Official documentation
Dikte: README (GitHub)
yusufipk, GitHub (yusufipk/dikte), 2026
Read the originalRead on October 22, 2026
- What it is
- The README of Dikte, a dictation app for Linux, macOS and Windows, in the maker's own GitHub repository. It describes where speech to text and cleanup run by default, how the result is pasted, and the cleanup rules for names.
- How much weight it carries
- The maker's own repository page, read on 2026-10-22. It is the maker's own description, not a test. The README says speech runs on this machine by default but also says cleanup uses a hosted provider by default, so the rows state both.
Facts taken from this source
- Dikte's README says the names listed under Cleanup rules go to the transcription model as a hint and to the cleanup model as a glossary.
“The names you list under Cleanup rules go to the transcription model as a hint and to the cleanup model as a glossary” (What it does section, Misheard words are repaired)
- Dikte's README says a cleanup model is asked to fix misheard words from context.
“so the cleanup model is asked to fix those from context” (What it does section, Misheard words are repaired)
- Dikte's README says cleanup uses a hosted provider by default, so transcript text goes to a third-party AI unless a local model is chosen.
“cleanup uses a hosted provider by default, but can also run on a local model” (Install section, last paragraph before the provider table)
- Dikte's README says speech to text runs locally by default.
“Speech to text runs locally by default; cleanup uses a hosted provider by default” (Install section, last paragraph before the provider table)
- Dikte's README says the result is pasted into whatever window you were typing in.
“the result lands in your clipboard and is pasted into whatever window you were typing in” (Opening section, first paragraph)