Skip to main content

kapi exec voice-infer

kapi voice-infer [files...]

Infer a draft voice profile from a content corpus using an LLM provider

Commandexec voice-infer
Try it (demo)

Demo mode — illustrative output from a built-in stub, not a real model. Install the CLI to run with your own API key.

kapi exec voice-infer
Flags
FlagTypeDefaultDescription
--api-keystringAPI key for the AI provider
--concurrency, -jint0max parallel files (0 = auto)
--credentialstringsaved credential name to use (see 'kapi credentials list')
--domainstringSubject domain hint for the analysis (e.g. medical legal technology)
--encoding, -estringcharset the input is read in; a byte-order mark in the file wins (default: the project's defaults.encoding, then UTF-8)
--enginestringformat engine for detected files: native or a plugin name (default: native, then plugins)
--fail-on-unknownboolfalseexit with error if any file cannot be processed (default: skip with warning)
--format, -fstringoverride input format detection
--jsonboolfalseoutput results as JSON
--map, -mstringArraymap glob pattern to format (e.g. '*.docx=okf_openxml:test')
--max-examplesint3Maximum before/after examples to infer
--modelstringAI model name
--no-warnboolfalsesuppress warnings for skipped files
--packboolfalsewhen transforming a .kpz, also eject the result to the .kpz (auto-pack)
--parallel-blocksint0fan out block processing across N goroutines (0 = off)
--profile-namestringName for the inferred draft profile
--progressboolfalseshow progress bar
--project, -pstringpath to a kapi.yaml project recipe or its directory (auto-discovered from cwd if omitted)
--providerstringanthropicAI provider (Anthropic, OpenAI, Gemini, Azure OpenAI, Claude Code, Ollama, Demo (illustrative))
--source-langstringsource language (e.g. en, en-US; defaults to the project's source_language, else en)
--strictboolfalsealias for --fail-on-unknown
--tracestringwrite flow trace JSON to file (for flow visualization)

← Back to the Command Reference