Assemble & split

The page-level toolkit lives under /v1/pdf/*. Every operation takes PDFs as base64 strings in a JSON body and — unless noted — returns the resulting PDF binary directly.

# Encode a file once, reuse it in the requests below
PDF=$(base64 -w0 input.pdf)     # macOS: base64 -i input.pdf

Merge

POST /v1/pdf/merge — combine 2+ PDFs, in order:

curl -X POST https://api.pdf-forge.dev/v1/pdf/merge \
  -H "Authorization: Bearer $PDFFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"pdfs\": [\"$(base64 -w0 a.pdf)\", \"$(base64 -w0 b.pdf)\"]}" \
  --output merged.pdf

import { readFile, writeFile } from "node:fs/promises";

const a = new Uint8Array(await readFile("a.pdf"));
const b = new Uint8Array(await readFile("b.pdf"));

const merged = await client.merge([a, b]);
await writeFile("merged.pdf", merged.buffer);

Split

POST /v1/pdf/split — one PDF per page. This one returns JSON:

curl -X POST https://api.pdf-forge.dev/v1/pdf/split \
  -H "Authorization: Bearer $PDFFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"pdf\": \"$PDF\"}"

{ "pages": ["<base64 page 1>", "<base64 page 2>"], "page_count": 2,
  "message": "PDF split into 2 pages" }

const result = await client.split(pdfBytes);
// result.pages: Uint8Array[], result.pageCount: number

Extract, remove, reorder

Page selections are arrays of strings: single pages ("3") and ranges ("5-8").

# Keep only these pages, in the given order
curl -X POST https://api.pdf-forge.dev/v1/pdf/extract-pages \
  -H "Authorization: Bearer $PDFFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"pdf\": \"$PDF\", \"pages\": [\"1\", \"3\", \"5-8\"]}" \
  --output extracted.pdf

# Delete these pages, keep the rest
curl -X POST https://api.pdf-forge.dev/v1/pdf/remove-pages \
  -H "Authorization: Bearer $PDFFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"pdf\": \"$PDF\", \"pages\": [\"2\", \"4\"]}" \
  --output trimmed.pdf

Reordering = extract-pages

extract-pages preserves the order you give. To reorder a 3-page document, extract all pages in the new sequence: "pages": ["3", "1", "2"].

Rotate

POST /v1/pdf/rotate — rotation must be a non-zero multiple of 90 (negative for counter-clockwise). Optional pages limits the scope:

curl -X POST https://api.pdf-forge.dev/v1/pdf/rotate \
  -H "Authorization: Bearer $PDFFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d "{\"pdf\": \"$PDF\", \"rotation\": 90, \"pages\": [\"1\"]}" \
  --output rotated.pdf

One unit per operation

Each successful call meters one unit against your monthly quota, like a generation. Errors are not counted.

Next steps

Compress — shrink the merged result.
Secure — protect, watermark, number the pages.
PDF & images — rasterize pages or build a PDF from scans.

Merge

Split

Extract, remove, reorder

Rotate

Next steps

On this page