File inspection
PDF metadata viewer and JavaScript check.
Open a PDF to read who made it and when, how many times it was saved, whether it is encrypted, and whether its structure contains JavaScript, auto-run actions, launch actions or embedded files. A keyword scan, not an antivirus, and nothing is uploaded.
Look inside a PDF before you open it
A PDF is more than pages. It is a structured file of numbered objects that can hold a document information dictionary (title, author, creator, producer, dates), a separate XMP metadata packet, form fields, links, attachments, encryption settings, and actions: things the reader does on its own, such as running JavaScript or jumping somewhere when the file opens. This inspector reads that structure straight from the file bytes, in your browser, and reports what it finds. It never renders the pages and it never executes anything in the file.
People use it to check what a PDF reveals about who made it before they send it on, to see whether a PDF that arrived by email has script or launch actions, to find out whether a file was edited and re-saved after it was first created, and to answer "check PDF for JavaScript" without sending the document to a third-party site.
What the report contains
- Information dictionary. Title, Author, Subject, Keywords, Creator (the authoring program), Producer (the program that wrote the PDF), CreationDate and ModDate. Dates are shown in readable form with the original string alongside.
- XMP metadata. The newer, XML-based metadata packet: creator, creator tool, producer, create, modify and metadata dates, and document and instance IDs. Many PDFs carry both, and the two can disagree.
- Version and structure. The version in the header (for example 1.7), whether the file is linearized, and how many cross-reference sections it has. More than one section means the file was saved incrementally: edits were appended to the end instead of rewriting the file, so older content can still be present.
- Encryption. Whether an
/Encryptdictionary exists, and its security handler revision and cipher family. Many encrypted PDFs open without a password because the user password is empty; the encryption mainly enforces permission flags such as no printing or copying. - Active-content keywords. The names that mark script and actions:
/JSand/JavaScript,/OpenAction,/AA(additional actions),/Launch,/EmbeddedFile,/URI,/SubmitForm,/AcroForm,/XFA,/RichMedia,/3Dand/JBIG2Decode, each with an explanation of why it matters.
How the scan works
- The file is read into memory and the header, trailer and cross-reference markers are located. Files over 50 MB are refused with a clear message. At most 64 compressed object streams are decoded, only when they use a single Flate filter and never beyond 64 MB of decompressed data in total; filter chains and nested attachments are not unpacked. A 10-second time limit stops runaway files.
- Binary stream bodies are cut out, so compressed picture data cannot create false keyword hits.
- Compressed object streams, which modern PDFs use to store most of the dictionaries, are inflated with the browser's
DecompressionStream(capped at 32 MB each) and scanned too. Without this step, script inside an object stream would be invisible. - Names written with
#xxhex escapes, such as/J#61vaScript, are decoded before matching. The PDF specification allows that spelling and malware has used it to dodge simple string searches. If any are found, the report says so. - The trailer's
/Inforeference is followed, to the object in the file or inside an object stream, and its entries are parsed. Plain-text JavaScript strings are quoted so you can read them.
Worked example
A "scanned contract" arrives from an unfamiliar sender. The report shows PDF 1.7, Producer "Microsoft Print to PDF", and flags: OpenAction 1, JavaScript 3, Launch 1. The quoted script is a short app.launchURL call. Scanned documents do not normally need any of these, so you do not open the file in a desktop reader and ask the sender to confirm. Contrast a government form downloaded from an agency: it shows AcroForm 1, JavaScript 12 and no Launch. Fillable forms use JavaScript for field calculations and formatting, so that result is expected. The same keyword can be harmless or harmful; context decides.
Common mistakes
- Treating "no JavaScript found" as "safe". A clean scan lowers one kind of risk. It says nothing about exploits against bugs in a PDF reader, malicious links, or social-engineering content on the pages.
- Treating "JavaScript found" as "malware". Fillable forms, calculators and signature workflows legitimately use it.
- Assuming metadata is trustworthy. Author and dates are whatever the producing program wrote and can be edited or forged. They are good for curiosity and for checking what you are about to share, not for proving authorship.
- Forgetting incremental saves. A "redacted" PDF saved as an update can still contain the original text earlier in the file. A count of cross-reference sections above 1 is a prompt to look closer.
- Expecting counts to equal links or scripts. Counts are keyword occurrences. One link normally shows
/URItwice, once as the action type and once as the address key.
What this tool does not do
- It is a keyword and structure scan, not an antivirus or a complete PDF parser. Damaged or unusual files, unsupported filters, and anything inside encrypted objects can be missed. For encrypted files the report says it cannot see inside and does not claim a clean result.
- It does not read page text, form values or embedded file contents, and does not extract attachments.
- It does not remove metadata or active content. To strip author and dates, use the PDF metadata remover; this tool shows you what is there first.
- It does not render or run the PDF, so opening an untrusted file here is no riskier than choosing it in a file picker.
Your PDF stays in your browser
The inspector runs inside an isolated frame with no network permission and no advertising or analytics code. The file is read from disk by your browser and scanned locally; the report is built in memory. Nothing is uploaded or stored, and the ads on this page, which sit outside the frame, cannot see which file you choose.
Common mix-ups
-
Often assumedPDF is an Adobe-only format, so its structure cannot be inspected without Adobe software.
What the source saysPDF 1.7 was published as the international standard ISO 32000-1 in July 2008, so its file structure is publicly specified. That is what lets a tool read a PDF's objects, metadata and scripts directly. [1]
Checked against these sources on .
Frequently asked questions
How do I check a PDF for JavaScript?
Choose the PDF in this tool. It scans the file structure, including compressed object streams, for /JS and /JavaScript names and lists how many it finds, with any plain-text script quoted. It runs in your browser and does not execute the script.
Does JavaScript in a PDF mean it is malicious?
No. Fillable forms use JavaScript for calculations and formatting. It is a reason to look closer when a document that should be static, such as a scan or invoice, contains script or auto-run actions.
Where is a PDF’s author and creation date stored?
In the document information dictionary (Author, CreationDate, ModDate) and often again in an XMP metadata packet. This tool reads both and shows them side by side. They are editable and not proof of authorship.
What does OpenAction mean?
An OpenAction runs when the document is opened. Most often it just goes to the first page, but it can also trigger JavaScript or another action, so it is flagged for review.
How can I tell whether a PDF was edited after it was created?
Look at the number of cross-reference sections and the creation and modification dates. More than one section means incremental saves, so earlier content may remain in the file. Dates alone can be forged.
Can it scan an encrypted PDF?
Only the unencrypted structure is visible. The report shows the encryption revision and cipher but cannot see strings or script inside encrypted objects, so it makes no clean claim for those files.
Is this an antivirus scanner?
No. It looks for known active-content keywords and structure. It cannot detect exploits against reader bugs or judge what a script does.
Is my PDF uploaded?
No. The file is scanned locally inside a frame with no network access and no ad or analytics code.
Sources
Sources last checked 2026-11-11.
- PDF 32000-1:2008 (PDF 1.7), Adobe-hosted copy: opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008.pdf (file structure, information dictionary, actions, encryption, name escapes).
- XMP Specification Part 1, Adobe: github.com/adobe/XMP-Toolkit-SDK/blob/main/docs/XMPSpecificationPart1.pdf (XMP data model and properties).
- DecompressionStream, MDN: developer.mozilla.org/en-US/docs/Web/API/DecompressionStream (in-browser inflate for object streams).
- Blob.slice, MDN: developer.mozilla.org/en-US/docs/Web/API/Blob/slice (reading files locally).
- PDF file format, Library of Congress description: www.loc.gov/preservation/digital/formats/fdd/fdd000277.shtml (format overview).