Command Reference
Every command in the kapi CLI, generated from the binary so the flags and synopses match the version you have installed. Select a command to read its description, flags, and examples. Commands that run entirely in the browser carry a Run button that executes them against a small sample file. AI and MT commands carry a Demo badge and run against a built-in stub, so their output is illustrative rather than from a real model. Commands that need network access, an API key, or local system access point to a walkthrough instead. Each command has a shareable link.
The groups are the ones the installed kapi prints in its help. In every project, kapi --help shows Work, led by look, check, fix, rules, connect, share and open, and Assets. Languages appears there when the recipe declares a target language, and kapi help languages lists it in any project. Advanced is listed under kapi help --all. Every command runs whether or not the default help lists it. Below, each group lists its commands alphabetically.
168 of 168 commands
Work
addRunAdd file patterns to the project's content10 flagsapplyRunApply a change set: content edits, review decisions, terms and recipe fields11 flagscheckRunVerify content against its rules and fail on a failing finding, like tests over code24 flagsconnectWatchPut an assistant on this project: rules as it starts, kapi check on every edit2 flagscontextRunAsk what applies before you write a file9 flagscontext atlasRunPrint the project's context records by axis and over time3 flagscontext logRunList what was recorded, or what this session recorded7 flagscontext noteRunRecord what you noticed, or what a person changed14 flagscontext placesRunLook up the places a geo axis takes2 flagscontext resetRunGo back to how the context stood at an earlier point4 flagscontext reviewRunDecide about the suggestions waiting for you13 flagscontext searchRunAsk what the project says about a word or phrase6 flagscontext suggestRunFile what kapi can see in the files as suggestions for you to decide3 flagscontext syncRunShare the context with the rest of the team7 flagsfixRunApply the fixes kapi can make on its own9 flagsinitRunScaffold a new kapi project in the current directory10 flagslookRunSee the project, or one file with its rules and findings in place1 flaglsRunList the files the project's content tracks8 flagsopenWatchOpen this folder in Kapi Desktop2 flagsrmRunRemove file patterns from the project's content6 flagsrulesRunDecide, list, narrow or drop the rules your project keeps8 flagsrules dropRunSet a suggestion or a kept rule aside2 flagsrules findRunFind the rules, suggestions and decisions that mention a word1 flagrules keepRunKeep a suggestion as a rule4 flagsrules listRunList the rules in force and the suggestions waiting1 flagrules narrowRunLimit a kept rule to one collection, named directly or by a folder3 flagsshareWatchShare the rules with your team8 flagsstatusRunShow how ready the project's content is against its gates8 flagsupRunBring the project up to date: run the default flow and stop where a gate holds content back37 flags
Assets
credentialsWatchManage saved AI provider credentialscredscredentials addWatchSave a named AI provider credential4 flagscredentials listWatchList saved credentialscredentials removeWatchRemove a saved credentialcredentials testWatchTest that a credential's API key is accessiblemodelsWatchManage attached LLMs and ML modelsmodels defaultWatchShow or set the default AI model (and its provider)6 flagsmodels listWatchList every model kapi can use (Ollama, plugin assets, cloud providers)1 flagmodels ollamaWatchManage the local Ollama runtime for on-device AI tools1 flagmodels ollama installWatchPrint how to install the Ollama runtime (or install it on macOS with --run)1 flagmodels ollama listWatchList models installed on the Ollama servermodels ollama pullWatchDownload a model onto the Ollama servermodels ollama statusWatchShow whether the Ollama runtime is installed, running, and which models are presentmodels pruneWatchRemove a model (Ollama model, or a cached plugin asset)models pullWatchDownload a model (Ollama model, or a plugin asset)models setupWatchInteractively pick and verify the default AI providertermsRunManage terminologyterms expandDemoPropose the surface forms each term takes (writes the terms)17 flagsterms exportRunExport terms to CSV, JSON, TBX, or a native .terms.json bundle9 flagsterms importRunImport terms from CSV, JSON, TBX, or a native .terms.json bundle12 flagsterms listRunList named terms stores in KAPI_HOMEterms lookupRunLook up a term in the terms store8 flagsterms occurrencesRunShow where a term is used in the project's content7 flagsterms searchRunSearch concepts in the terms store7 flagsterms statsRunShow terms statistics4 flagsterms validateRunCheck terms for structural problems and missing forms10 flagsvoiceRunKeep AI-generated content in voice (tone, style, terminology)voice checkDemoScore text against a voice profile16 flagsvoice editRunEdit a voice profile in your editor7 flagsvoice importRunImport a profile YAML into the local voice store8 flagsvoice newRunScaffold a voice profile YAML to fill in (optionally seeded from a starter pack)2 flagsvoice packRunInstall a built-in starter pack into the local voice store8 flagsvoice profilesRunList voice profiles (local store + built-in packs)8 flagsvoice rewriteDemoSubstitute forbidden/competitor terms for their approved wording (offline)12 flagsvoice showRunShow a voice profile as a guide15 flagsvoice validateRunValidate a voice profile YAML (structure, enums, regex, terms)5 flags
Languages
memoryRunManage content memorymemory auditRunTrace content-memory entries by merge batch id6 flagsmemory exportRunExport content memory to TMX or a native .memory.json bundle7 flagsmemory importRunImport a TMX file or a native .memory.json bundle into content memory9 flagsmemory import-dirRunImport all TMX files from a directory into content memory9 flagsmemory listRunList named Memories in KAPI_HOMEmemory lookupRunLook up text in content memory8 flagsmemory searchRunSearch content memory entries7 flagsmemory sessionsRunManage TMX import sessionsmemory sessions deleteRunRemove a session record (entries are retained, session_id cleared)3 flagsmemory sessions listRunList all import sessions3 flagsmemory sessions showRunShow details for a single import session3 flagsmemory statsRunShow content memory statistics4 flagspseudo-translateRunPseudo-translate files to test locale readiness, with no AI keys needed27 flagstranslateDemoTranslate files with guardrails: content memory reuse, AI translate, then checks27 flags
Advanced
completionRunGenerate shell completion scriptconfigRunGet or set kapi configuration (app config subcommands, recipe keys positionally)1 flagconfig getRunPrint a single config value5 flagsconfig listRunList every configured value, including plugin namespaces5 flagsconfig pathRunPrint the config file path (of kapi, or of a plugin namespace)5 flagsconfig setRunSet a config value (persists to the global config file, or a plugin's when the key is namespaced)5 flagsconfig unsetRunRemove a config value, restoring the built-in default5 flagsexecRunExecute one registry tool on files (the raw layer under the porcelain verbs)exec case-transformRunCase Transform22 flagsexec create-targetRunCreate Target23 flagsexec diff-leverageRunDiff Leverage21 flagsexec dnt-checkRunDo-Not-Translate Checkdnt19 flagsexec encoding-detectRunEncoding Detect15 flagsexec entity-extractDemoDetect named entities (people, organizations, products, locations) with an LLM, a local NER model, or both23 flagsexec external-commandRunExternal Command26 flagsexec inline-codes-removeRunInline Codes Remove24 flagsexec media-refineDemoRe-read low-confidence OCR/ASR lines with a configurable multimodal LLM23 flagsexec placeholder-checkRunPlaceholder Check18 flagsexec pseudo-translateRunPseudo Translatepseudo22 flagsexec qaDemoCheck translation quality with deterministic rules or an LLM (select with --mode)60 flagsexec recycleRunRecycle28 flagsexec redactRunReplace sensitive spans with protected placeholders before processing22 flagsexec remove-targetRunRemove Target23 flagsexec reviewDemoReview translations with scoring using an LLM provider24 flagsexec scriptRunScript22 flagsexec search-replaceRunSearch and Replace26 flagsexec segmentationDemoSplit source text into sentence or chunk segments (stand-off overlay)segment33 flagsexec term-checkRunTerminology Check20 flagsexec term-extractDemoExtract candidate terminology from content using an LLM provider20 flagsexec translateDemoTranslate content with an LLM or machine-translation provider (select an engine, then a provider)30 flagsexec translate-afterRunTranslate After16 flagsexec unredactRunRestore original values into redacted content after processing18 flagsexec voice-checkDemoCheck text against a voice profile using an LLM provider20 flagsexec voice-inferDemoInfer a draft voice profile from a content corpus using an LLM provider22 flagsexec whitespace-correctRunWhitespace Correct31 flagsexec xml-validationRunXML Validation20 flagsextractRunEmit a bilingual file for a translator: native .kpz or XLIFF/PO15 flagsflowsRunList available flowsflows listRunList available flowsformatsRunList supported file formats3 flagsformats explainRunSay which format reads a file, and why3 flagsformats infoRunShow detailed information about a formatformats listRunList supported file formats2 flagsformats schemaRunOutput JSON Schema for a formatinfoRunReport a project's packable state, or a .kpz archive's contents6 flagsinspectRunRead any format into content blocks, with the references and revisions an edit names7 flagsmcpWatchStart an MCP server (stdio) for assistants writing in this project5 flagsmergeRunApply a returned bilingual file (.kpz/XLIFF/PO) back onto the project source6 flagspackRunSnapshot a kapi project's working state into a .kpz9 flagspluginWatchInstall and manage manifest-driven pluginspluginsplugin doctorWatchRun health/self-checks on installed pluginsplugin infoWatchShow details for an installed pluginplugin installWatchInstall a plugin from the registry3 flagsplugin listWatchList installed pluginsplugin pruneWatchRemove retired plugins that are still installed2 flagsplugin rebuild-cacheWatchForce a rebuild of the plugin dispatch cacheplugin registryWatchManage plugin registriesplugin registry addWatchAdd a registry to global config1 flagplugin registry listWatchList configured registriesplugin registry removeWatchRemove a registry from global configplugin removeWatchRemove an installed pluginplugin searchWatchSearch the registry for plugins1 flagplugin updateWatchUpdate an installed plugin to the latest matching version4 flagsplugin update-indexWatchRefresh the cached registry index1 flagplugin verifyWatchRe-verify an installed plugin's manifest and binaryresolveRunRebase or discard a version of a .kpz document that did not land3 flagsrunRunRun a composed flow (a named multi-tool pipeline)27 flagsstatsRunSummarize content metrics for files: blocks, words, characters, segments, by role4 flagsstoreRunMaintain this project's context storestore exportRunWrite this project's context and its history to one file2 flagsstore importRunRead context files, or a file written by export, into this project2 flagsstore localesRunReport the locale each stored row is filed under1 flagstore rebuildRunRebuild this project's store from its history2 flagstelemetryWatchShow or change anonymous usage telemetry (status, on, off)telemetry offWatchDisable anonymous usage telemetry (persists to the global config file)telemetry onWatchEnable anonymous usage telemetry (persists to the global config file)telemetry statusWatchShow whether telemetry is enabled, why, and the exact payload shapetoolsRunList available processing toolstools listRunList available processing toolstools schemaRunPrint the JSON Schema for a tool's parameterstrustWatchList, inspect and change which projects may run commandstrust allowWatchRecord that a project may run the commands its recipe names1 flagtrust listWatchList every recorded decision and whether it still appliestrust revokeWatchWithdraw the decision recorded for a project1 flagtrust showWatchShow what a recipe would run and the decision that applies1 flagunpackRunRehydrate a project's working state from a .kpz snapshot6 flags