---
name: "pdf123-repair"
description: "Repair corrupted or damaged PDF files online with PDF123 - rebuild structure so the file opens again. Runs the PDF123 \"Repair PDF\" tool (pdf123.xyz) over its REST API with curl, no account needed. Use when the user wants this done to their file. Also known as: 修复 PDF, Reparar PDF, PDF सुधारें, إصلاح PDF, Perbaiki PDF, Réparer le PDF, Исправить PDF, PDFの修復, PDF reparieren, PDF 복구, Sửa PDF, PDF onar, 修復 PDF, Ripara PDF, ซ่อม PDF, Napraw PDF, Полагодити PDF, PDF repareren, Baiki PDF, Reparera PDF, Επισκευή PDF, Поправка на PDF, Repara PDF, Rekebisha PDF."
compatibility: "Needs curl 7.76+ and outbound HTTPS to pdf123.xyz, or PDFX_API_BASE pointing at a self-hosted pdfx-server."
---

# Repair PDF (PDF123)

Repair corrupted or damaged PDF files online with PDF123 - rebuild structure so the file opens again.

Web version: https://pdf123.xyz/repair · All tools: https://pdf123.xyz/skills/pdf123.md

## When to use

- Rebuild xref on an export another tool wrote poorly
- Try qpdf before giving up on a slightly damaged file
- Normalize structure before Merge

## Run it

Replace the sample file names and values with the user's, then run:

```bash
API="${PDFX_API_BASE:-https://pdf123.xyz}"
curl -sS --fail-with-body -X POST "$API/api/v1/misc/repair" \
  -F "fileInput=@input.pdf" \
  --output-dir "pdf123-output/$(date +%Y%m%d-%H%M%S)" --create-dirs -OJ -w '%{filename_effective} %{content_type}\n'
```

## Inputs

Everything is `multipart/form-data`. The command above already sends each field with its default; keep them all and change only the values the user asked for, since some endpoints reject a missing optional field.

| Field | Type | Required | Default | Notes |
| --- | --- | --- | --- | --- |
| `fileInput` | file | yes | | .pdf (one file) |

## Result

curl saves the result in a new `pdf123-output/<timestamp>/` directory under the server's file name and prints its path and content type. Several output files come back as one ZIP. Tell the user where the file is.

## Limits

- No deep object reconstruction
- Not Unlock
- qpdf-unreadable files fail
- Upload limit on this website: 500 MB per file, sent in chunks above about 95 MB. A single direct API request body is capped at 100 MB.

## Errors

- A non-zero curl exit means the request failed. The saved file then holds `application/problem+json`; read it and report its `detail` to the user instead of retrying blindly.
- `413`: the upload exceeds 100 MiB. `429`: wait for `Retry-After` seconds, then retry once.
- Send `X-API-KEY: $PDFX_API_KEY` only if the user has a PDF123 API key; anonymous calls work without it.

## Privacy

Files are uploaded to the API host, processed, and deleted once the response is sent. For confidential files, ask before uploading, or use a self-hosted server via `PDFX_API_BASE`.