Commands

A full list of the CLI command and its respective help text.

configure

Usage: genaiscript configure [options]

Interactive help to configure providers

Options:
  -p, --provider <string>  Preferred LLM provider aliases (choices: "openai",
                           "azure", "azure_serverless",
                           "azure_serverless_models", "github", "ollama",
                           "anthropic", "anthropic_bedrock", "google",
                           "huggingface", "mistral", "alibaba", "deepseek",
                           "transformers", "lmstudio", "jan", "llamafile",
                           "litellm", "whisperasr")
  -h, --help               display help for command

run

Usage: genaiscript run [options] <script> [files...]

Runs a GenAIScript against files.

Options:
  -p, --provider <string>                    Preferred LLM provider aliases (choices: "openai", "azure", "azure_serverless", "azure_serverless_models", "github", "ollama", "anthropic", "anthropic_bedrock", "google", "huggingface", "mistral", "alibaba", "deepseek", "transformers", "lmstudio", "jan", "llamafile", "litellm", "whisperasr")
  -m, --model <string>                       'large' model alias (default)
  -sm, --small-model <string>                'small' alias model
  -vm, --vision-model <string>               'vision' alias model
  -ma, --model-alias <nameid...>             model alias as name=modelid
  -re, --reasoning-effort <string>           Reasoning effort for o* models (choices: "high", "medium", "low")
  -lp, --logprobs                            enable reporting token probabilities
  -tlp, --top-logprobs <number>              number of top logprobs (1 to 5)
  -ef, --excluded-files <string...>          excluded files
  -egi, --exclude-git-ignore                 exclude files that are ignored through the .gitignore file in the workspace root
  -ft, --fallback-tools                      Enable prompt-based tools instead of builtin LLM tool calling builtin tool calls
  -o, --out <string>                         output folder. Extra markdown fields for output and trace will also be generated
  -rmo, --remove-out                         remove output folder if it exists
  -ot, --out-trace <string>                  output file for trace
  -oo, --out-output <string>                 output file for output
  -od, --out-data <string>                   output file for data (.jsonl/ndjson will be aggregated). JSON schema information and validation will be included if available.
  -oa, --out-annotations <string>            output file for annotations (.csv will be rendered as csv, .jsonl/ndjson will be aggregated)
  -ocl, --out-changelog <string>             output file for changelogs
  -pr, --pull-request <number>               pull request identifier
  -prc, --pull-request-comment [string]      create comment on a pull request with a unique id (defaults to script id)
  -prd, --pull-request-description [string]  create comment on a pull request description with a unique id (defaults to script id)
  -prr, --pull-request-reviews               create pull request reviews from annotations
  -tm, --teams-message                       Posts a message to the teams channel
  -j, --json                                 emit full JSON response to output
  -y, --yaml                                 emit full YAML response to output
  -fe, --fail-on-errors                      fails on detected annotation error
  -r, --retry <number>                       number of retries (default: "10")
  -rd, --retry-delay <number>                minimum delay between retries (default: "1000")
  -md, --max-delay <number>                  maximum delay between retries (default: "10000")
  -l, --label <string>                       label for the run
  -t, --temperature <number>                 temperature for the run
  -tp, --top-p <number>                      top-p for the run
  -mt, --max-tokens <number>                 maximum completion tokens for the run
  -mdr, --max-data-repairs <number>          maximum data repairs
  -mtc, --max-tool-calls <number>            maximum tool calls for the run
  -se, --seed <number>                       seed for the run
  -em, --embeddings-model <string>           embeddings model for the run
  -c, --cache                                enable LLM result cache
  -cn, --cache-name <name>                   custom cache file name
  -cs, --csv-separator <string>              csv separator (default: "\t")
  -ff, --fence-format <string>               fence format (choices: "xml", "markdown", "none")
  -ae, --apply-edits                         apply file edits
  --vars <namevalue...>                      variables, as name=value, stored in env.vars. Use environment variables GENAISCRIPT_VAR_name=value to pass variable through the environment
  -rr, --run-retry <number>                  number of retries for the entire run
  --no-run-trace                             disable automatic trace generation
  -h, --help                                 display help for command

test

Usage: genaiscript test|eval [options] [command]

Options:
  -h, --help                 display help for command

Commands:
  run [options] [script...]  Runs the tests for scripts
  list [options]             List available tests in workspace
  view                       Launch test viewer
  help [command]             display help for command

test run

Usage: genaiscript test run [options] [script...]

Runs the tests for scripts

Arguments:
  script                              Script ids. If not provided, all scripts
                                      are tested

Options:
  --redteam                           run red team tests
  -p, --provider <string>             Preferred LLM provider aliases (choices:
                                      "openai", "azure", "azure_serverless",
                                      "azure_serverless_models", "github",
                                      "ollama", "anthropic",
                                      "anthropic_bedrock", "google",
                                      "huggingface", "mistral", "alibaba",
                                      "deepseek", "transformers", "lmstudio",
                                      "jan", "llamafile", "litellm",
                                      "whisperasr")
  -m, --model <string>                'large' model alias (default)
  -sm, --small-model <string>         'small' alias model
  -vm, --vision-model <string>        'vision' alias model
  -ma, --model-alias <nameid...>      model alias as name=modelid
  -re, --reasoning-effort <string>    Reasoning effort for o* models (choices:
                                      "high", "medium", "low")
  --models <models...>                models to test where mode is the key
                                      value pair list of m (model), s (small
                                      model), t (temperature), p (top-p)
  -o, --out <folder>                  output folder
  -rmo, --remove-out                  remove output folder if it exists
  --cli <string>                      override path to the cli
  -td, --test-delay <string>          delay between tests in seconds
  --cache                             enable LLM result cache
  -v, --verbose                       verbose output
  -pv, --promptfoo-version [version]  promptfoo version, default is 0.104.0
  -os, --out-summary <file>           append output summary in file
  -g, --groups <groups...>            groups to include or exclude. Use :!
                                      prefix to exclude
  -h, --help                          display help for command

test list

Usage: genaiscript test list [options]

List available tests in workspace

Options:
  --redteam                 list red team tests
  -g, --groups <groups...>  groups to include or exclude. Use :! prefix to
                            exclude
  -h, --help                display help for command

test view

Usage: genaiscript test view [options]

Launch test viewer

Options:
  -h, --help  display help for command

convert

Usage: genaiscript convert [options] <script> [files...]

Converts file through a GenAIScript. Each file is processed separately through
the GenAIScript and the LLM output is saved to a <filename>.genai.md (or custom
suffix).

Options:
  -s, --suffix <string>              suffix for converted files
  -rw, --rewrite                     rewrite input file with output (overrides
                                     suffix)
  -cw, --cancel-word <string>        cancel word which allows the LLM to notify
                                     to ignore output
  -ef, --excluded-files <string...>  excluded files
  -egi, --exclude-git-ignore         exclude files that are ignored through the
                                     .gitignore file in the workspace root
  -m, --model <string>               'large' model alias (default)
  -sm, --small-model <string>        'small' alias model
  -vm, --vision-model <string>       'vision' alias model
  -ma, --model-alias <nameid...>     model alias as name=modelid
  -ft, --fallback-tools              Enable prompt-based tools instead of
                                     builtin LLM tool calling builtin tool
                                     calls
  -o, --out <string>                 output folder. Extra markdown fields for
                                     output and trace will also be generated
  --vars <namevalue...>              variables, as name=value, stored in
                                     env.vars. Use environment variables
                                     GENAISCRIPT_VAR_name=value to pass
                                     variable through the environment
  -c, --cache                        enable LLM result cache
  -cn, --cache-name <name>           custom cache file name
  -cc, --concurrency <number>        number of concurrent conversions
  -h, --help                         display help for command

scripts

Usage: genaiscript scripts|script [options] [command]

Utility tasks for scripts

Options:
  -h, --help                  display help for command

Commands:
  list [options] [script...]  List all available scripts in workspace
  create <name>               Create a new script
  fix                         fix all definition files
  compile [folders...]        Compile all scripts in workspace
  model [options] [script]    List model connection information for scripts
  help [command]              display help for command

scripts list

Usage: genaiscript scripts list [options] [script...]

List all available scripts in workspace

Arguments:
  script                    Script ids

Options:
  -g, --groups <groups...>  groups to include or exclude. Use :! prefix to
                            exclude
  -h, --help                display help for command

scripts create

Usage: genaiscript scripts create [options] <name>

Create a new script

Arguments:
  name        Name of the script

Options:
  -h, --help  display help for command

scripts fix

Usage: genaiscript scripts fix [options]

fix all definition files

Options:
  -h, --help  display help for command

scripts compile

Usage: genaiscript scripts compile [options] [folders...]

Compile all scripts in workspace

Arguments:
  folders     Pattern to match files

Options:
  -h, --help  display help for command

scripts model

Usage: genaiscript scripts model [options] [script]

List model connection information for scripts

Arguments:
  script       Script id or file

Options:
  -t, --token  show token
  -h, --help   display help for command

cache

Usage: genaiscript cache [options] [command]

Cache management

Options:
  -h, --help      display help for command

Commands:
  clear [name]    Clear cache
  help [command]  display help for command

cache clear

Usage: genaiscript cache clear [options] [name]

Clear cache

Arguments:
  name        Name of the cache, tests

Options:
  -h, --help  display help for command

video

Usage: genaiscript video [options] [command]

Video tasks

Options:
  -h, --help                       display help for command

Commands:
  probe <file>                     Probes metadata from a video/audio file
  extract-audio [options] <file>   Transcode video/audio file
  extract-frames [options] <file>  Extract video frames
  help [command]                   display help for command

video probe

Usage: genaiscript video probe [options] <file>

Probes metadata from a video/audio file

Arguments:
  file        Audio or video file to inspect

Options:
  -h, --help  display help for command

video extract-audio

Usage: genaiscript video extract-audio [options] <file>

Transcode video/audio file

Arguments:
  file                 Audio or video file to transcode

Options:
  -t, --transcription  Convert audio for speech-to-text
  -h, --help           display help for command

video extract-frames

Usage: genaiscript video extract-frames [options] <file>

Extract video frames

Arguments:
  file                             Audio or video file to transcode

Options:
  -k, --keyframes                  Extract only keyframes (intra frames)
  -st, --scene-threshold <number>  Extract frames with a minimum threshold
  -c, --count <number>             maximum number of frames to extract
  -s, --size <string>              size of the output frames wxh
  -f, --format <string>            Image file format
  -h, --help                       display help for command

retrieval

Usage: genaiscript retrieval|retreival [options] [command]

RAG support

Options:
  -h, --help                           display help for command

Commands:
  search [options] <query> [files...]  Search using vector embeddings
                                       similarity
  fuzz [options] <query> [files...]    Search using string distance
  help [command]                       display help for command

retrieval search

Usage: genaiscript retrieval search [options] <query> [files...]

Search using vector embeddings similarity

Options:
  -ef, --excluded-files <string...>  excluded files
  -tk, --top-k <number>              maximum number of results
  -h, --help                         display help for command

retrieval fuzz

Usage: genaiscript retrieval fuzz [options] <query> [files...]

Search using string distance

Options:
  -ef, --excluded-files <string...>  excluded files
  -tk, --top-k <number>              maximum number of results
  -h, --help                         display help for command

serve

Usage: genaiscript serve [options]

Start a GenAIScript local server

Options:
  -p, --port <number>       Specify the port number, default: 8003
  -k, --api-key <string>    API key to authenticate requests
  -n, --network             Opens server on 0.0.0.0 to make it accessible on
                            the network
  -c, --cors <string>       Enable CORS and sets the allowed origin. Use '*' to
                            allow any origin.
  --remote <string>         Remote repository URL to serve
  --remote-branch <string>  Branch to serve from the remote
  --remote-force            Force pull from remote repository
  --remote-install          Install dependencies from remote repository
  --dispatch-progress       Dispatch progress events to all clients
  -h, --help                display help for command

parse

Usage: genaiscript parse|parsers [options] [command] <file...>

Parse various outputs

Arguments:
  file                         input JSONL files

Options:
  -h, --help                   display help for command

Commands:
  data [options] <file>        Convert CSV, YAML, TOML, INI, XLSX, XML, MD/X
                               frontmatter or JSON data files into various
                               formats
  fence <language> <file>      Extracts a code fenced regions of the given type
  pdf [options] <file>         Parse a PDF into text and images
  docx [options] <file>        Parse a DOCX into texts
  html-to-text <file>          Parse an HTML file into text
  code <file> [query]          Parse code using tree sitter and executes a
                               query
  tokens [options] <files...>  Count tokens in a set of files
  jsonl2json                   Converts JSONL files to a JSON file
  prompty [options] <file...>  Converts .prompty files to genaiscript
  jinja2 [options] <file>      Renders Jinj2 or prompty template
  secrets <file...>            Applies secret scanning and redaction to files

parse data

Usage: genaiscript parse data [options] <file>

Convert CSV, YAML, TOML, INI, XLSX, XML, MD/X frontmatter or JSON data files
into various formats

Options:
  -f, --format <string>  output format (choices: "json", "json5", "yaml",
                         "ini", "csv", "md")
  -h, --help             display help for command

parse fence

Usage: genaiscript parse fence [options] <language> <file>

Extracts a code fenced regions of the given type

Options:
  -h, --help  display help for command

parse pdf

Usage: genaiscript parse pdf [options] <file>

Parse a PDF into text and images

Options:
  -i, --images        extract images
  -o, --out <string>  output folder
  -h, --help          display help for command

parse docx

Usage: genaiscript parse docx [options] <file>

Parse a DOCX into texts

Options:
  -f, --format <string>  output format (choices: "markdown", "html", "text")
  -h, --help             display help for command

parse html-to-text

Usage: genaiscript parse html-to-text [options] <file>

Parse an HTML file into text

Options:
  -h, --help  display help for command

parse code

Usage: genaiscript parse code [options] <file> [query]

Parse code using tree sitter and executes a query

Options:
  -h, --help  display help for command

parse tokens

Usage: genaiscript parse tokens [options] <files...>

Count tokens in a set of files

Options:
  -ef, --excluded-files <string...>  excluded files
  -h, --help                         display help for command

parse jsonl2json

Usage: genaiscript parse jsonl2json [options]

Converts JSONL files to a JSON file

Options:
  -h, --help  display help for command

parse prompty

Usage: genaiscript parse prompty [options] <file...>

Converts .prompty files to genaiscript

Arguments:
  file                input JSONL files

Options:
  -o, --out <string>  output folder
  -h, --help          display help for command

parse jinja2

Usage: genaiscript parse jinja2 [options] <file>

Renders Jinj2 or prompty template

Arguments:
  file                   input Jinja2 or prompty template file

Options:
  --vars <namevalue...>  variables, as name=value passed to the template
  -h, --help             display help for command

parse secrets

Usage: genaiscript parse secrets [options] <file...>

Applies secret scanning and redaction to files

Arguments:
  file        input files

Options:
  -h, --help  display help for command

info

Usage: genaiscript info [options] [command]

Utility tasks

Options:
  -h, --help                display help for command

Commands:
  help                      Show help for all commands
  system                    Show system information
  env [options] [provider]  Show .env information

info help

Usage: genaiscript info help [options]

Show help for all commands

Options:
  -h, --help  display help for command

info system

Usage: genaiscript info system [options]

Show system information

Options:
  -h, --help  display help for command

info env

Usage: genaiscript info env [options] [provider]

Show .env information

Options:
  -t, --token   show token
  -e, --error   show errors
  -m, --models  show models if possible
  -h, --help    display help for command

models

Usage: genaiscript models [options] [command]

Options:
  -h, --help       display help for command

Commands:
  list [provider]  List all available models
  alias            Show model alias mapping
  help [command]   display help for command

models list

Usage: genaiscript models list [options] [provider]

List all available models

Options:
  -h, --help  display help for command

models alias

Usage: genaiscript models alias [options]

Show model alias mapping

Options:
  -h, --help  display help for command