This version of the Speechly documentation is no longer actively maintained. For up-to-date documentation, see the latest version.

Speechly CLI Reference

List of all Speechly CLI commands and their usage.

annotate

Create SAL annotations for a list of examples using Speechly.

speechly annotate [<input file>] [<app id>] [flags]

Examples

speechly annotate -a <app_id> --input input.txt
speechly annotate -a <app_id> --input input.txt > output.txt
speechly annotate -a <app_id> --reference-date 2021-01-20 --input input.txt > output.txt

To evaluate already deployed Speechly app, you need a set of evaluation examples that users of your application might say.

Options

  -a, --app string              App ID of the application to evaluate. Can alternatively be given as the first positional argument.
  -d, --de-annotate             Instead of adding annotation, remove annotations from output.
  -e, --evaluate                Print evaluation stats instead of the annotated output.
  -h, --help                    help for annotate
  -i, --input string            Evaluation utterances, separated by newline, if not provided, read from stdin. Can alternatively be given as the first positional argument.
  -o, --output string           Where to store annotated utterances, if not provided, print to stdout.
  -r, --reference-date string   Reference date in YYYY-MM-DD format, if not provided use current date.

convert

Converts an Alexa Interaction Model in JSON format to a Speechly configuration

speechly convert [-l language] <input_file> [flags]

Examples

speechly convert my-alexa-skill.json
speechly convert -l en-US my-alexa-skill.json

Options

  -h, --help              help for convert
  -l, --language string   Language of input (default "en-US")

create

Create a new application in the current project

speechly create [<application name>] [flags]

Options

  -h, --help              help for create
  -l, --language string   Application language. Current only 'en-US' and 'fi-FI' are supported. (default "en-US")
  -n, --name string       Application name, can alternatively be given as the sole positional argument.

delete

Delete an existing application

speechly delete [<app_id>] [flags]

Options

  -a, --app string   Application ID to delete. Can alternatively be given as the sole positional argument.
  -d, --dry-run      Don't perform the deletion.
  -f, --force        Skip confirmation prompt.
  -h, --help         help for delete

deploy

Send the contents of a local directory to training

Synopsis

The contents of the directory given as argument is sent to the API and validated. Then, a new model is trained and automatically deployed as the active model for the application.

speechly deploy [<app_id>] <directory> [flags]

Examples

speechly deploy <app_id> /path/to/config
speechly deploy -a <app_id> .

Options

  -a, --app string        Application to deploy the files to. Can be given as the first positional argument.
  -h, --help              help for deploy
      --skip-validation   Skip the validation step. If there are validation issues, they will not be shown, the deploy will fail silently.
  -w, --watch             Wait for training to be finished.

describe

Print details about an application

speechly describe [<app_id>] [flags]

Options

  -a, --app string   Application ID to describe. Can alternatively be given as the sole positional argument.
  -h, --help         help for describe
  -w, --watch        If app status is training, wait until it is finished.

download

Download the active configuration or model bundle of the given app.

Synopsis

Fetches the currently stored configuration or model bundle. This command does not check for validity of the stored configuration, but downloads the latest version.

speechly download [<app_id>] <directory> [flags]

Examples

speechly download <app_id> /path/to/config
speechly download -a <app_id> .
speechly download -a <app_id> . --model tflite

Options

  -a, --app string     Application which configuration or model bundle to download. Can be given as the first positional argument.
  -h, --help           help for download
  -m, --model string   Model bundle machine learning framework. Available options are: ort, tflite, coreml and all. This feature is available on Enterprise plans (https://speechly.com/pricing)

edit

Edit an existing application

speechly edit [flags]

Options

  -a, --app string    Application ID
  -h, --help          help for edit
  -n, --name string   Application name

evaluate

Compute accuracy between annotated examples (given by ‘speechly annotate’) and ground truth.

speechly evaluate [<app_id>] [<input_file>] [flags]

Examples

speechly evaluate --input output.txt --ground-truth ground-truth.txt

Options

  -t, --ground-truth string   Manually verified ground-truths for annotated examples.
  -h, --help                  help for evaluate
  -i, --input string          SAL annotated utterances, as given by 'speechly annotate' command.

list

List applications in the current project

speechly list [flags]

Options

  -h, --help   help for list

projects

Manage API access to Speechly projects

speechly projects [flags]

Options

  -h, --help   help for projects

projects add

Add access to a pre-existing project

speechly projects add [apikey] [flags]

Options

      --apikey string            API token, created in Speechly Dashboard. Can also be given as the sole positional argument.
  -h, --help                     help for add
      --host string              API address (default "api.speechly.com")
      --name string              An unique name for the project. If not given the project name configured in Speechly Dashboard will be used.
      --skip-online-validation   Skips validating the API token against the host.

projects remove

Remove access to a project

speechly projects remove [flags]

Options

  -h, --help          help for remove
      --name string   The name for the project for which access is to be removed.

projects use

Select the default project used

speechly projects use [flags]

Options

  -h, --help          help for use
      --name string   An unique name for the project.

sample

Sample a set of examples from the given SAL configuration

Synopsis

The contents of the directory given as argument is sent to the API and compiled. If configuration is valid, a set of examples are printed to stdout.

speechly sample [<app_id>] <directory> [flags]

Examples

speechly sample -a <app_id> .
speechly sample -a <app_id> /path/to/config
speechly sample <app_id> /path/to/config --stats

Options

  -a, --app string                 Application to sample the files from. Can alternatively be given as the first positional argument.
      --batch-size int             How many examples to return. Must be between 32 and 10000. (default 100)
      --seed int                   Random seed to use when initializing the sampler.
      --stats                      Print intent and entity distributions to the output.
      --advanced-stats             Print entity type, value and value pair distributions to the output.
      --advanced-stats-limit int   Line limit for advanced_stats. The lines are ordered by count. (default 10)
  -h, --help                       help for sample

stats

Get utterance statistics for the current project or an application in it

speechly stats [<app_id>] [flags]

Examples

speechly stats [<app_id>]
speechly stats -a APP_ID
speechly stats > output.csv
speechly stats --start-date 2021-03-01 --end-date 2021-04-01

Options

  -a, --app string          Application to get the statistics for. Can alternatively be given as the sole positional argument.
      --end-date string     End date for statistics, not included in results.
      --export              Print report as CSV
  -h, --help                help for stats
      --start-date string   Start date for statistics.

transcribe

Transcribe the given jsonlines file

speechly transcribe <input_file> [flags]

Examples

speechly transcribe <input_file>

Options

  -h, --help           help for transcribe
  -m, --model string   On-device model bundle file

utterances

Get a sample of recent utterances.

Synopsis

Fetches a sample of recent utterances and their SAL-annotated transcript.

speechly utterances <app_id> [flags]

Options

  -h, --help   help for utterances

validate

Validate the given configuration for syntax errors

Synopsis

The contents of the directory given as argument is sent to the API and validated. Possible errors are printed to stdout.

speechly validate [<app_id>] <directory> [flags]

Examples

speechly validate -a <app_id> .
speechly validate <app_id> /path/to/config

Options

  -a, --app string   Application to validate the files for. Can alternatively be given as the first positional argument.
  -h, --help         help for validate

version

Print the version number

speechly version [flags]

Options

  -h, --help   help for version

Last updated by Mathias Lindholm on October 4, 2022 at 16:47 +0300

Found an error on our documentation? Please file an issue or make a pull request