A CLI for Infinite Red, with functionality we use internally.
Alias: t
Transcribes the given audio file and writes the transcription to a text file with the same name.
This uses OpenAI's whisper-1
model.
We primarily use this for transcribing React Native Radio's podcast episodes.
--format
: The response format you want the transcription in (default is 'json').--prompt
: An optional prompt to include in the transcription request.
OPENAI_API_KEY
: You must set this environment variable with your OpenAI API key.
The filename is required. All --options are optional.
# can also be set in .zshrc file or similar
# windows powershell is $env:OPENAI_API_KEY = 'yourkeyhere'
export OPENAI_API_KEY=yourkeyhere
# transcribe a file
npx infinitered transcribe myfile.mp3 --format=json --prompt="Optional prompt" --output=mytranscription.txt
Make sure you have set the OPENAI_API_KEY
in your environment before running the command.
Clone it down:
git clone git@github.com:infinitered/infinitered-cli.git
Install dependencies:
cd infinitered-cli
yarn
Link the CLI:
yarn link
Now you can run the CLI from anywhere:
ir transcribe <filename>
# or
infinitered transcribe <filename>
infinitered-cli is powered by Gluegun. See Gluegun's docs for more information on how to use it.
To package your CLI up for NPM, do this:
$ yarn login
$ yarn whoami
$ yarn test
$ yarn build
$ yarn publish
MIT - see LICENSE