CLI reference — pdfstudio

The pdfstudio CLI reuses the exact same content-stream engine as the web app — true in-place text editing, burned-in redaction, and page operations. 100% local: no network calls, document bytes never leave the machine.

Setup

git clone https://github.com/Kayforkind/NavigatorsLab-PDF-Studio
cd NavigatorsLab-PDF-Studio
npm install
npm run build:packages

Run from the repo root with npx pdfstudio …, or make it global and call pdfstudio anywhere:

npm link -w packages/cli

Conventions

• - as an input reads the PDF from stdin.
• -o - writes binary output to stdout (pipe-friendly). PDF Studio refuses to write binary PDF to a terminal — pipe it or use -o <file>.
• Exit codes: 0 ok · 1 error · 2 no matches / nothing changed.
• --json on read commands (info, extract-text, search) emits machine-readable output for scripting.
• Human-readable reports go to stderr; data goes to stdout.

Commands

info — metadata and page geometry

pdfstudio info contract.pdf
pdfstudio info contract.pdf --json

extract-text — text per page

pdfstudio extract-text contract.pdf -p 1-3
pdfstudio extract-text contract.pdf -o text.txt
cat in.pdf | pdfstudio extract-text - --json

Flag	Description
-p, --pages <spec>	1-based pages, e.g. "1-3,5"
-o, --output <file>	Write to file (- = stdout)
--json	Machine-readable output

search — find text

pdfstudio search contract.pdf "termination clause"
pdfstudio search contract.pdf "Termination" --case-sensitive --json

Prints page/line/snippet hits. Exits 2 with no matches on stderr when nothing is found.

edit-text — true in-place text editing

pdfstudio edit-text in.pdf --find "Acme Corp" --replace "Globex Inc" --all -o out.pdf
pdfstudio edit-text in.pdf --find "draft" --replace "" -o out.pdf   # delete
cat in.pdf | pdfstudio edit-text - --find "draft" --replace "FINAL" -o - > out.pdf

The original bytes are deleted from the content stream, not overlaid — same engine as the app's Edit tool.

Flag	Description
--find <text>	Required. Text to find
--replace <text>	Required. Replacement ("" deletes)
-p, --pages <spec>	1-based page ranges, e.g. "1-3,5" (default: all)
--all	Replace every occurrence (default: first match only)
--case-sensitive	Case-sensitive match
-o, --output <file>	Required. Output PDF (- = stdout)

redact — burned-in redaction

pdfstudio redact in.pdf --find "123-45-6789" -o clean.pdf
pdfstudio redact in.pdf --rect "2:72,500,200,24" -o clean.pdf
pdfstudio redact in.pdf --rect "1:72,700,300,20" --rect "3:72,700,300,20" -o clean.pdf

Covered text is deleted from the file — unrecoverable. --rect takes <page:x,y,w,h> in PDF points, origin at the bottom-left of the page's media box (repeatable). --find redacts every line containing the text.

Flag	Description
--find <text>	Redact every line containing this text
--rect <spec>	<page:x,y,w,h> in points (repeatable)
-p, --pages <spec>	Limit --find to these 1-based pages
--case-sensitive	Case-sensitive --find
-o, --output <file>	Required. Output PDF (- = stdout)

merge — combine PDFs

pdfstudio merge a.pdf b.pdf -o combined.pdf

Merges inputs in order. - may be used for one of the inputs (stdin).

split — one PDF per range

pdfstudio split in.pdf --ranges 1-3 --ranges 4-6 -o "part-%d.pdf"

--ranges is repeatable (1-based, e.g. "1-3"). The output template supports %d (part index) and %s (sanitized range spec). Use -o - for a single range to stdout.

rotate — rotate pages

pdfstudio rotate in.pdf --angle 90 -p 1-2 -o rotated.pdf

Flag	Description
--angle <deg>	Required. 90, 180, or 270 (clockwise)
-p, --pages <spec>	1-based pages (default: all)
-o, --output <file>	Required. Output PDF (- = stdout)

pages — delete and reorder

pdfstudio pages in.pdf --delete 5 --order 3,1,2 -o reordered.pdf

Flag	Description
--delete <ranges>	1-based pages to drop, e.g. "2,5-7"
--order <list>	1-based new order, e.g. "3,1,2" — unlisted pages keep relative order, appended
-o, --output <file>	Required. Output PDF (- = stdout)

Privacy guarantee for scripts

The CLI makes zero network calls. Parsing (pdf.js), editing (content-stream rewriting), and writing (pdf-lib) all run in process. Redaction deletes the text operators from the file, and the output is rebuilt so orphaned bytes are gone too.

Prompt-injection note

PDF content is untrusted input. If you feed CLI output into an agent or script that treats text as instructions, keep the boundary: document text is data, never instructions. (The MCP server wraps extraction results in explicit delimiters — see MCP server.)