Skip to content

File formats

What exav does with each format it knows, built from the project’s format registry (formats/registry.json, which the tests check against the code):

  • Malware scan. Every file is scanned for signatures as bytes, whatever its format. A container is opened and every member is scanned too, some documents are also scanned through the text they show, and a format exav recognises but does not open is reported as not examined, never as clean. The scanner’s formats page has the detail.
  • Extract. What exav-unpack takes out: the members of an archive, or the parts of a document (the embedded files and streams of a PDF, the XML parts of a DOCX). Streamed means decoded as it is read, at any size; read whole means held in memory first. See the extractor’s formats.
  • Extractor code. The module of exav-unpack that reads the format, and the crates its Cargo feature brings in.
  • View. Whether the viewer draws it in the browser, and a file of that format to open in the demo when it has one.
126 formats
FormatExtensionsMalware scanExtractExtractor codeViewDemoComment
Archives and compressed streams
ZIP.zipOpened, every member scannedIts members, streamedzip.rsdeflate64flate2zipYesOpen
Packages that are ZIP files (JAR, APK, IPA, wheel, NuGet, VSIX, MSIX...).jar.war.ear.apk.apks.xapk.aab.ipa.xpi.whl.nupkg.vsix.kmz.appx.appxbundle.msix.msixbundleOpened, every member scannedIts members, streamedzip.rsdeflate64flate2zipYesOpen
First part of a split archive (.7z.001, .zip.001, .rar.001, .tar.001, .z01).7z.001.zip.001.rar.001.tar.001.z01Opened, every member scannedIts members, streamedzip.rsdeflate64flate2zipYesOpened with the other parts when the files around it have them.
Comic book archives (CBZ, CBR, CB7, CBT).cbz.cbr.cb7.cbtOpened, every member scannedIts members, streamedzip.rsdeflate64flate2zipYesOpenA ZIP, RAR, 7z or tar under another name.
gzip.gz.tgzOpened, every member scannedIts members, streamedgzip.rsflate2YesOpen
bzip2.bz2.tbz.tbz2Opened, every member scannedIts members, streamedbzip2.rscrc32fasttinyvecYesOpen
xz.xz.txzOpened, every member scannedIts members, streamedxz.rscrc32fastxz4rustYesOpen
Zstandard.zst.tzstOpened, every member scannedIts members, streamedzstd.rsruzstdYesOpen
LZ4.lz4Opened, every member scannedIts members, streamedlz4.rsYes
LZMA (.lzma).lzmaOpened, every member scannedIts members, streamedlzmaalone.rslzma-rust2YesLZMA-alone, as LZMA Utils wrote it. No magic number: typed by the header's own rules (properties, dictionary size, size, first range-coder byte).
RPM package.rpmOpened, every member scannedIts members, streamedrpm.rsYesThe install scripts and the payload (a compressed cpio) are members.
lzip.lz.tlzOpened, every member scannedIts members, streamedlzip.rscrc32fastlzma-rust2Yes
Unix compress (.Z).z.tazOpened, every member scannedIts members, streamedlzw.rsYes
tar.tarOpened, every member scannedIts members, streamedtar.rsYesOpen
7z.7zOpened, every member scannedIts members, streamedsevenzcrc32fastflate2lzma-rust2tinyvecYesOpen
RAR (RAR3 and RAR5).rarOpened, every member scannedIts members, streamedrar.rsYesOpen
Cabinet (CAB).cabOpened, every member scannedIts members, streamedcab.rsflate2lzxdYesOpen
Compiled HTML Help (CHM).chmOpened, every member scannedIts members, streamedchm.rslzxdYes
LHA / LZH.lzh.lhaOpened, every member scannedIts members, streamedlha.rsdelharcYes
ARJ.arjOpened, every member scannedIts members, read wholearj.rsbitstream-iocrc32fastdelharcYes
ar (static libraries, Debian packages).ar.a.deb.udebOpened, every member scannedIts members, streamedar.rsYesOpen
cpio (RPM payloads, initramfs).cpioOpened, every member scannedIts members, streamedcpio.rsYesOpen
xar (macOS packages).xar.pkg.xipOpened, every member scannedIts members, streamedxar.rscrc32fastflate2lzma-rust2tinyvecxz4rustYesMembers stored or in the zlib, bzip2 and LZMA styles are decoded as they are read. A member in another style is scanned as stored and reported.
ZOO.zooOpened, every member scannedIts members, read wholezoo.rsdelharcYes
ARC / PKARC / PAK.arcOpened, every member scannedIts members, read wholearc.rsYes
ALZ.alzOpened, every member scannedIts members, read wholealz.rscrc32fastflate2tinyvecYes
EGG.eggOpened, every member scannedIts members, read wholeegg.rscrc32fastflate2lzma-rust2tinyvecYes
InstallShield .zOpened, every member scannedIts members, read wholeishield_z.rsNo
MS-Compress (SZDD, KWAJ)Opened, every member scannedIts members, streamedszdd.rsNo
ACE.aceRecognised, reported as not examinedRecognised, reportedace.rsNoRecognised, never decoded: no trustworthy implementation to validate a decoder against.
StuffIt and StuffIt X.sit.sitxRecognised, reported as not examinedRecognised, reportedstuffit.rsNo
InstallShield InstallScript cabinetOpened, every member scannedIts members, streamedishield_cab.rsflate2NoA cabinet holding its header is opened. Files in the other cabinets of a set (data1.hdr, data2.cab...) are reported.
CryptFFOpened, every member scannedIts members, streamedcryptff.rsNoA 16-byte header, then the content with every byte flipped: one member, decoded as it is read. There is no key.
lrzip.lrzRecognised, reported as not examinedRecognised, reportedNo
AppleSingle and AppleDoubleOpened, every member scannedIts members, streamedapplesingle.rsNo
Disk images and filesystems
ISO 9660 and UDF.isoOpened, every member scannedIts members, streamediso.rsYesOpen
Apple disk image (DMG).dmgOpened, every member scannedIts members, streameddmg.rsbase64crc32fastflate2quick-xmltinyvecxz4rustYes
VHD.vhdOpened, every member scannedIts members, streamedvhd.rsYes
VHDX.vhdxOpened, every member scannedIts members, streamedvhdx.rsflate2Yes
VMDK (and OVA).vmdk.ovaOpened, every member scannedIts members, streamedvmdk.rsflate2Yes
QCOW2.qcow2Opened, every member scannedIts members, streamedqcow2.rsflate2Yes
Windows imaging (WIM, ESD).wim.esdOpened, every member scannedIts members, streamedwimsha1Yes
Partition maps (GPT, APM, MBR) and raw disk images.imgOpened, every member scannedIts members, streamedpartition.rsYes
FAT12, FAT16, FAT32Opened, every member scannedIts members, streamedfat.rsNo
NTFSOpened, every member scannedIts members, streamedntfs.rslznt1No
ext2, ext3, ext4Opened, every member scannedIts members, streamedext.rsNo
exFATOpened, every member scannedIts members, streamedexfat.rsNo
SquashFS.squashfs.sqfs.snapOpened, every member scannedIts members, streamedsquashfs.rscrc32fastflate2lzma-rust2ruzstdxz4rustNoBoth byte orders; gzip, LZMA, LZO, XZ, LZ4 and Zstandard blocks.
cramfs.cramfsOpened, every member scannedIts members, streamedcramfs.rsflate2No
romfs.romfsOpened, every member scannedIts members, streamedromfs.rsNo
EROFS.erofsOpened, every member scannedIts members, streamederofs.rsflate2lzma-rust2ruzstdNoFiles stored plain, in chunks or compressed (LZ4, LZMA, DEFLATE, Zstandard) are read; a file whose tail is in the packed inode is reported.
Installers, self-extractors and packers
InstallShield MSIOpened, every member scannedIts members, streamedishield_msi.rsflate2NoThe setup executable of InstallShield 2009 to 2011, and of 2015 and later: the installer MSI and the files that go with it. The second is opened by default and not under --clamav-compat. A header of another type is reported.
NSIS installerOpened, every member scannedIts members, read wholensis.rscrc32fastflate2lzma-rust2tinyvecNo
Inno Setup installerOpened, every member scannedIts members, streamedinno.rscrc32fastflate2lzma-rust2tinyvecNoVersions 5.1.5 to 6.4 with the files in the executable. Other versions, encrypted files and files in a separate setup-1.bin are reported.
Self-extracting archivesOpened, every member scannedIts members, streamedsfx.rscrc32fastNo
AutoIt compiled scriptsOpened, every member scannedIts members, read wholeautoit.rsNo
UPX-packed executablesOpened, every member scannedIts members, read wholeupx.rsbase64flate2lzma-rust2No
Other packed PE executablesOpened, every member scannedIts members, read wholepepack.rsNoUnpacked by native decoders, or by running the stub in a bounded emulator.
Mach-O universal binariesOpened, every member scannedIts members, streamedmachofat.rsNo
Documents
OLE2 compound files (legacy Office, MSI)Opened, every member scannedIts parts, streamedole.rsencoding_rsflate2No
PDF.pdfOpened, every member scannedIts parts, read wholepdf.rsflate2YesOpen
Word (DOCX).docx.docm.dotxOpened, every member scannedIts parts, streamedzip.rsdeflate64flate2zipYesOpen
Excel (XLSX).xlsx.xlsmOpened, every member scannedIts parts, streamedzip.rsdeflate64flate2zipYesOpen
PowerPoint (PPTX).pptx.pptm.ppsx.potxOpened, every member scannedIts parts, streamedzip.rsdeflate64flate2zipYesOpen
Word 6 to 2003 (DOC).doc.dotScanned with its textIts parts, streamedole.rsencoding_rsflate2YesOpenThe scanner reads its text with visible-text-office.
Excel 4 to 2003 (XLS).xls.xltScanned with its textIts parts, streamedole.rsencoding_rsflate2YesOpenThe scanner reads its text with visible-text-office.
PowerPoint 95 to 2003 (PPT).ppt.pps.potScanned with its textIts parts, streamedole.rsencoding_rsflate2YesOpenPowerPoint 2 to 4 are recognised and reported unsupported in the viewer.
RTF.rtfScanned with its textIts parts, read wholertf.rsYesOpen
OpenDocument text (ODT).odt.ott.fodtScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
OpenDocument spreadsheet (ODS).ods.ots.fodsScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
OpenDocument presentation (ODP).odp.otp.fodpScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
Keynote.keyScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
Numbers.numbersScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
Pages.pagesScanned with its textIts parts, streamedzip.rsdeflate64flate2zipNoConverted by exav-render, but the viewer has no matcher for it yet.
CSV and TSV.csv.tsvScanned with its textNoYesOpen
Visio.vsd.vst.vdx.vtx.vsdx.vsdm.vstx.vstmScanned as bytesNoYesOpenVSDX, VSDM, VSTX and VSTM by the package's converter; the binary and XML ones by the external converter.
CorelDRAW.cdr.cmxScanned as bytesNoYesBy the external converter.
Publisher.pubOpened, every member scannedIts parts, streamedole.rsencoding_rsflate2YesBy the external converter.
PageMaker 6 to 7.pmd.pm6.p65Scanned as bytesNoYesBy the external converter.
FreeHand 3 to 11.fh.fh3.fh4.fh5.fh7.fh8.fh9.fh10.fh11Scanned as bytesNoYesBy the external converter.
QuarkXPress 3.1 to 4.1.qxd.qxtScanned as bytesNoYesBy the external converter.
Zoner Draw 4 and 5.zmfScanned as bytesNoYesBy the external converter.
WordPerfect Graphics 1 and 2.wpgScanned as bytesNoYesBy the external converter.
Hangul Word Processor 3 (HWP3).hwpOpened, every member scannedIts parts, read wholehwp3.rsflate2No
OneNote.oneOpened, every member scannedIts parts, streamedonenote.rsNo
Adobe XDP.xdpOpened, every member scannedIts parts, read wholexdp.rsbase64No
Mail and encodings
TNEF (winmail.dat)winmail.dat.tnefOpened, every member scannedIts members, streamedtnef.rsYes
Email (MIME) and MHTML.eml.mht.mhtmlOpened, every member scannedIts members, streamedemail_stream.rsbase64mail-parserNoA file that starts with one of 19 header names (`From `, `From: `, `Received: `, `To: `, `Subject: `, `Date: ` and others, spelt as ClamAV types them) or whose header block has a MIME header. A message is also read from inside a text file or binary data that holds two of its header lines close together.
BinHex.hqxOpened, every member scannedIts members, read wholebinhex.rsNo
uuencode.uuOpened, every member scannedIts members, read wholeuuencode.rsbase64No
Images
PNG, JPEG, GIF, WebP and SVG.png.jpg.jpeg.webp.gif.svgTyped, signatures appliedNoYesOpenDrawn by the browser; SVG in an <img>, which runs no script.
TIFF.tif.tiffTyped, signatures appliedNoYesOpen
BMP.bmpTyped, signatures appliedNoYesOpen
ICO, PNM, QOI, DDS, farbfeld, Radiance HDR.ico.cur.pbm.pgm.ppm.pnm.pam.qoi.dds.ff.hdrTyped, signatures appliedNoYesOpen
Targa (TGA).tga.icbScanned as bytesNoYesNo magic number: the viewer reads a file as Targa when its header is sane and its pixels fill it.
JPEG 2000.jp2.jpf.jpx.j2k.j2cTyped, signatures appliedNoYesOpen
JBIG2.jb2.jbig2Typed, signatures appliedNoYesOpen
JPEG XL.jxlTyped, signatures appliedNoYesOpen
Camera RAW (DNG, CR2, CR3, NEF, ARW, ORF, RW2, RAF, PEF...).dng.gpr.cr2.cr3.crw.nef.nrw.arw.sr2.srf.orf.ori.rw2.rwl.raw.raf.pef.srw.3fr.fff.iiq.x3f.mrw.dcr.kdc.erf.mef.mosTyped, signatures appliedNoYesOpenShown by the preview the file embeds; the sensor data is not decoded.
HEIC and HEIF.heic.heifTyped, signatures appliedNoNoTyped by the viewer so that it is not taken for a video; no plugin opens it.
Drawings and 3D models
OpenDocument drawing (ODG).odg.otg.fodgScanned with its textIts parts, streamedzip.rsdeflate64flate2zipYesOpen
AutoCAD DWG.dwgOpened, every member scannedIts parts, read wholedwgencoding_rsYesOpenR13 to R2018 are drawn; older releases are reported.
AutoCAD DXF.dxfOpened, every member scannedIts parts, read wholedxfencoding_rsYesOpen
IFC.ifcScanned as bytesNoYesOpen
STL.stlScanned as bytesNoYesOpen
Wavefront OBJ.objScanned as bytesNoYesOpen
Stanford PLY.plyScanned as bytesNoYesOpen
glTF 2.0 (GLTF, GLB).gltf.glbScanned as bytesNoYesOpen
Video and audio
Video (MP4, MOV, WebM, Ogg).mp4.m4v.mov.webm.ogvScanned as bytesNoYesOpenPlayed by the browser, so only what the browser decodes.
Audio (MP3, M4A, AAC, WAV, Ogg, Opus, FLAC).mp3.m4a.aac.wav.ogg.oga.opus.flacScanned as bytesNoYesOpen
Text and code
Markdown.md.markdownScanned with its textNoYesOpen
JSON.json.geojsonTyped, signatures appliedNoYesOpen
XML (and RSS, KML, GPX, plist...).xml.xsd.xsl.xslt.rss.atom.kml.gpx.rdf.wsdl.plistTyped, signatures appliedNoYesOpen
HTML.html.htm.xhtmlScanned with its textNoYesOpenShown as its source.
Diffs and patches.diff.patchTyped, signatures appliedNoYesOpen
Plain text, logs, configuration, subtitles.txt.text.log.jsonl.ndjson.srt.vtt.yaml.yml.toml.ini.cfg.conf.propertiesTyped, signatures appliedNoYesOpen
Source code.js.mjs.cjs.jsx.ts.tsx.mts.cts.css.scss.py.sh.bash.zsh.bat.ps1.c.h.cc.cpp.hpp.cs.rs.go.java.kt.swift.rb.php.pl.lua.sqlTyped, signatures appliedNoYesOpenScripts also get JavaScript views.
Executables and other carriers
Windows executables (PE).exe.dll.sysTyped, signatures appliedNoNo
ELFTyped, signatures appliedNoNo
Mach-OTyped, signatures appliedNoNo
Windows shortcuts (LNK).lnkOpened, every member scannedIts members, read wholelnk.rsNo
Python bytecode (.pyc).pycOpened, every member scannedIts members, streamedpyc.rsNo
Java class files.classOpened, every member scannedIts members, read wholejavaclass.rsNo
SWF (Flash).swfOpened, every member scannedIts members, streamedswf.rsflate2lzma-rust2No
AI models (pickle, safetensors).pkl.safetensorsOpened, every member scannedIts members, read wholeaimodel.rsNo
Microsoft Script Encoder.vbe.jseOpened, every member scannedIts members, read wholescrenc.rsNo