vikingowl/polyscribe
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[docs] update README with new CLI options, usage tips, and guidance for language models
Some checks failed
CI / ci (push) Has been cancelled
Details

Browse Source
This commit is contained in:
vikingowl
2025-08-12 05:17:41 +02:00
parent a26eade80b
commit af473c4942
1 changed files with 6 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
README.md
View File

@@ -30,8 +30,12 @@ Quickstart
- ./target/release/polyscribe --download-models - ./target/release/polyscribe --download-models
2) Transcribe a file: 2) Transcribe a file:
- ./target/release/polyscribe -v -o output my_audio.mp3 - ./target/release/polyscribe -v -o output --out-format json --jobs 4 my_audio.mp3
This writes JSON and SRT into the output directory with a date prefix. This writes JSON (because of --out-format json) into the output directory with a date prefix. Omit --out-format to write all available formats (JSON and SRT). For large batches, add --continue-on-error to skip bad files and keep going.
Gotchas
- English-only models: If you picked an English-only Whisper model (e.g., tiny.en, base.en), non-English language hints (via --language) will be rejected and detection may be biased toward English. Use a multilingual model (without the .en suffix) for non-English audio.
- Language hints help: When you know the language, pass --language <code> (e.g., --language de) to improve accuracy and speed. If the audio is mixed language, omit the hint to let the model detect.
Shell completions and man page Shell completions and man page
- Completions: ./target/release/polyscribe completions <bash|zsh|fish|powershell|elvish> > polyscribe.<ext> - Completions: ./target/release/polyscribe completions <bash|zsh|fish|powershell|elvish> > polyscribe.<ext>