kapi inspect

$ kapi inspect [flags] [FILE...]

Parse each file into one record per content block: the text, a stable content-hash anchor, and the block's structural role (heading, list-item, table-cell, …) and nesting level. Any format, whether a Word document, a JSON catalog, Markdown or HTML, yields the same shape, so an AI agent or RAG pipeline can read content, retrieve against the anchors, and write edits back to the same blocks.

Prints a JSON array by default; --output-format yaml emits a YAML sequence, and --jsonl streams one JSON object per line (JSONL) for piping into an ingestion pipeline.

Positional paths take glob patterns (** recurses) and directories, expanded by kapi itself. Inside a project, no FILE means the project's tracked content; FILE "-" reads standard input.

Commandinspect

Groupadvanced

Try it

Runs in your browser against a small sample file. Edit the command before running, or press Run to execute it as shown.

$ kapi inspect report.docx

Flags

Flag	Type	Default	Description
--encoding	string		input encoding
--format, -f	string		input format (default: auto-detect by extension/content)
--jsonl	bool	false	stream one JSON object per line (JSONL) instead of a JSON array
--project	stringSlice		also render each block to these target formats (html, markdown, asciidoc) under "projected"
--source-lang	string		source language (e.g. en, en-US; defaults to the project's source_language, else en)

More examples

$ kapi inspect --jsonl 'docs/**/*.md' | jq .content_hash

$ kapi inspect --output-format yaml report.dclg.xml

$ kapi inspect report.docx --project html,markdown # each block rendered per format

cat page.html | kapi inspect -f html

← Back to the Command Reference