rpptx-cli 0.14.1

CLI tool for inspecting, converting, and manipulating PPTX files
rpptx-cli-0.14.1 is not a library.
Visit the last successful build: rpptx-cli-0.0.0

rpptx-cli

rpptx-cli makes complete PPTX workflows available to shell scripts. It inspects and extracts deck content, validates package invariants, compares or replaces text, and produces deterministic fixed output.

Capabilities

  • Human-readable inspection, and JSON inspection with per-shape kind, placeholder, geometry, rotation, autofit, and text detail.
  • Slide-order text extraction and recursive outlines, as plain text or schema-1 JSON, with speaker notes on request.
  • PDF, PNG, JPEG, and multi-page TIFF conversion.
  • Selected-slide rendering, thumbnails, text diff, replacement, and validation.
  • Modern comment thread listing, addition, replies, resolution, and removal.
  • Scriptable output covers slide order, notes, comments, recursive group content, relationship-backed media, and deterministic rendering diagnostics.

Measured footprint and speed

Measurement Value Version Platform Build mode Input Command Statistic Measured on
Crates.io archive: rpptx-cli 62,359 compressed bytes, 287,346 member bytes, 8 members 0.14.1 macOS 26.6.2, Apple M5 Max, arm64 cargo publish --workspace --dry-run --allow-dirty --locked Tracked rpptx-cli package inventory python3 scripts/readme_doctests.py --record-measurements gzip archive bytes, tar member bytes, tar member count 2026-10-14

Use it when

Use the CLI for shell automation. Use rpptx when the same operations belong inside a Rust application.

Relationship

It uses oxml-cli-support for shared command conventions and the real rpptx facade for document behavior.

Example

cargo install rpptx-cli --version '^0.14.1'
rpptx inspect deck.pptx --json
rpptx text deck.pptx --json
rpptx outline deck.pptx --notes
rpptx comment add deck.pptx --slide 2 --author Reviewer \
  --text "Check the figures" --date 2026-09-25T10:00:00Z -o reviewed.pptx
rpptx comment list reviewed.pptx --json
rpptx convert deck.pptx --to pdf -o deck.pdf
rpptx thumbnail deck.pptx -o thumbnail.png

inspect --json keeps its existing keys and adds shape_details beside each slide's shape count. Each shape reports its z-order index, id, name, kind, placeholder type and index, direct position and size in EMU, rotation in degrees, direct autofit mode, paragraphs, table size, and children. A placeholder that inherits its geometry from its layout reports null geometry, and a placeholder without an explicit type reports a null type.

Both inspection forms report every core property the deck sets. Beside title, creator, subject, description, keywords, last modified by, created and modified, they report category, content status, identifier, language, last printed, revision and version, which inspect --json adds to its metadata object. Human-readable inspection prints (none) when the deck sets no core property.

text --json reports each slide's one-based number, slide id, paragraphs, and speaker notes, or null notes when the slide has no notes part. Each paragraph has a typed zero-based path of shape, table row, table cell, and paragraph positions, the id of the shape that owns it, its level, its visible text, and its regular runs. Run indexes are the facade's run positions, so fields and line breaks appear only in the paragraph text, where a line break is U+000B. Run formatting is null when a run has no direct properties. Otherwise it records nullable direct bold, italic, underline, font, point size, and sRGB colour values.

outline --json reports each slide's title, its outline items with their levels, and its speaker notes. Notes are the plain text of the notes body, with paragraphs and line breaks both written as newlines. --notes prints one Notes: line for each non-empty notes line after a slide's plain text or outline. JSON output always carries the notes.

comment works on modern PowerPoint comments. Legacy comments stay preserved but are not listed. list shows each comment and reply as open, resolved, or closed, and its JSON carries the resolved flag beside the raw status token. add --slide takes a one-based slide number. add and reply require an RFC 3339 --date, reuse an existing author with the same name, and add the author otherwise. New ids are sequential GUIDs, so the output depends on neither a clock nor a random source. resolve takes a thread id. remove takes a thread id, which also removes its replies, or a reply id. Every mutation requires -o/--output, refuses an existing output, publishes only a complete presentation, and supports a schema-1 record through --json.

convert, render, and thumbnail refuse an output file that already exists unless --force is given. Even with --force they refuse their own input file under any spelling of its path, and an output that is not a regular file, such as a directory, a symbolic link, a FIFO, or a device like /dev/null. A run checks every file it would write before it writes the first one, and publishes each file only once it is complete, so a failed run leaves no truncated output.

The output extension of replace and of every comment mutation selects the package class the output declares. A .potx template written to deck.pptx becomes a presentation, and a deck that carries a VBA project cannot change to a macro-free extension.

Run rpptx --help or rpptx <command> --help for the complete command surface.

One-shot editing and fit reports

rpptx replace deck.pptx --map pairs.json --json -o filled.pptx
rpptx slide add deck.pptx --layout 7 --at 1 --json -o added.pptx
rpptx slide duplicate deck.pptx 2 -o copied.pptx
rpptx slide remove deck.pptx 2 -o removed.pptx
rpptx slide move deck.pptx 3 --to 1 -o reordered.pptx
rpptx slide hide deck.pptx 2 -o hidden.pptx
rpptx slide show hidden.pptx 2 -o shown.pptx
rpptx notes set deck.pptx 1 --from-file notes.txt -o noted.pptx
rpptx fit deck.pptx --json
rpptx meta get deck.pptx --json
rpptx meta set deck.pptx --title "Quarterly report" --author Ada -o titled.pptx

Replacement maps are ordered JSON arrays of objects with placeholder, value and optional expect. Unknown fields and invalid types refuse. An absent or null expectation permits any count, including zero. Each pair searches the preceding pair's result, so repeated pairs and cascades are meaningful. An empty array succeeds unchanged. Empty placeholders refuse. Single -p/-v and --map are mutually exclusive, and map mode refuses a command-line --expect. Single replacement retains its refusal on zero matches unless --expect 0 is supplied. Slides and speaker notes share the native literal scope. All pairs run in one checked native candidate. A later count mismatch or malformed input creates no output and leaves the source unchanged.

Replacement JSON has schema1, action, mode, ordered pairs with index, placeholder, value, count and nullable expected, total_count and output. Every mutation prints JSON only after publication. New editing commands require an output, refuse an existing destination and protect input aliases, symlinks and hard links. They use the native extension-based package class and VBA contracts. Recoverable staging failures remove this invocation's new files. Filesystem cleanup remains best effort rather than crash-atomic publication.

Slide and numeric layout positions are one-based. A decimal layout selects its checked numeric index. Other layout values require an exact unique name. --at is the final insertion position, including append, and defaults to append. Creation and placement form one shared native transaction. Duplicate inserts directly after its source. Move uses the final destination position, and moving to the same position succeeds unchanged. Hide and show also retain no-op behavior. Mutation JSON records action, selected slide, resulting slide count, output and changed where applicable.

Notes require exactly one of --text and --from-file. Complete UTF-8 input is read before editing. LF, CRLF, carriage returns, leading/trailing spaces, Unicode and empty text are retained without trimming or an added newline. Existing notes body formatting and opaque payload stay preserved. JSON reports the selected slide and Unicode scalar character_count.

fit is read-only and consumes the native all-owner fit_report, including its inherited contexts, admitted table cells and diagnostics. It uses bundled deterministic fonts or the strict caller loader through --font-dir. Frames carry owner part, child path, cell, shape identity/name, context, mode, status, current stored values, overflow, needed total scale/reduction and the provider's quantized and persisted trials. Dimensions/heights are points, old/new bounds are EMU, and needed scale/reduction use hundred-thousandths. No permitted fitting trial means nullable needed_scale with reason=unfit-at-floor. Empty frames remain accounted for. Unsupported mandatory owners refuse with diagnostics instead of disappearing. Exit0 means all admitted frames were assessed without overflow, exit1 means overflow, and exit2 means a checked error. The command does not modify autofit or save.

Core metadata get includes every native core field. Set accepts title, author, subject, keywords, description and category, preserving all unspecified fields and opaque imported metadata. An empty supplied value removes that field. Duplicate flags and an empty set request refuse before publication.

The same common operations are exposed in Python as Presentation.replace_all, SlideCollection.add_slide_at and Presentation.fit_report. Ordered counts are tuples, count mismatches retain index/placeholder/expected/found, and failed or actual no-op batches preserve held content handles. FitReport, FitFrameReport and FitTrial are frozen snapshots rather than mutable owner selectors.

Licensed font embedding

Use caller-owned font files only after confirming their embedding license. Embedding requires explicit authorization and exact license identity. Commands write a new output and refuse input/font aliases or existing destinations.

rpptx font embed input.pptx --name "My Face" --font ./MyFace-Regular.ttf \
  --license-authorized --license-identity "SIL Open Font License 1.1" \
  --output embedded.pptx
rpptx font list embedded.pptx --json
rpptx font settings embedded.pptx --embed-true-type-fonts false --output disabled.pptx
rpptx font remove embedded.pptx --name "My Face" --kind regular --output removed.pptx

The four CLI kinds are regular, bold, italic and bold-italic. Settings accept true, false or none, with none removing the owned value. Listing reports actual native part, format and byte-length facts in the schema1 JSON envelope. PowerPoint preserves compressed/unknown imported font data and reports absent decoded bytes. save-subset-fonts is the other presence-sensitive setting.