filegrail

Local file provenance and metadata analysis.

Python 3.10+ 68 formats Runtime dependencies Local and read-only Network requests License

What it does · Evidence sources · Formats · Install · Usage · Examples

filegrail helps determine where a file came from, what it contains about its own history, and what happened to it on the machine.

It combines two sources of information:

Traces stored by the system and applications - browser download history, OS origin metadata, shell history, archives, torrents, recent-file records, sync folders and trash information. These traces can reveal where a file was downloaded from, which application handled it, when it appeared, where it was stored or whether it came from an archive or torrent.
Metadata stored inside the file - EXIF, XMP, IPTC, C2PA, document properties, media tags, email headers and other embedded data. Depending on the format, this can reveal the device, software, author, editor, timestamps, GPS coordinates, document history and other details.

File provenance is the whole subject. Under it, every record filegrail produces belongs to one of three categories, and the report keeps them apart:

Category	Question	Examples
Origin	How or from where did the file reach this environment?	Browser download history, `Zone.Identifier`, macOS where-from, XDG attributes, a fetch command, a `yt-dlp` sidecar
Metadata	What does the file record about itself?	EXIF, XMP, IPTC, document properties, media tags, Content Credentials
Activity	What happened to the file here?	Recent Documents, Windows shortcuts, trash records, sync folders, filesystem times

Each record also carries how it was matched to that file - a recorded path, a file name, a name and exact size, membership of a container, or the file's own bytes. A record tied to a file by nothing but its name is reported as exactly that.

filegrail does not say "acquisition" for a download. In digital forensics that word means the examiner taking custody of material - disk imaging, memory capture, a forensic copy - and it is not reused here for something else.

Scanning is local, read-only and makes no network requests. filegrail clean is the only command that writes files, and it writes cleaned copies to a separate directory.

What it does

Give filegrail a file or directory and it checks all supported provenance traces and embedded metadata that apply. Results always show where each piece of information came from.

If two independent sources disagree - for example, browser history and an OS origin attribute point to different URLs - both are shown and the conflict is reported instead of being silently resolved.

Capability	What you get
File provenance	Where a file came from and when: the address it was fetched from, the page that linked to it, the program that fetched it, the archive or torrent it came in - and which source said so
Metadata extraction	What a file records about its own earlier life: camera and body serial, where and when it was taken, author and software, revision count - decoded field by field
Evidence correlation	Agreement and conflicts between independent sources
Timeline	Origin, creation, editing and activity events in chronological order
Identifiers	Six types of pivot, each with the file, source and place it came from
File relationships	XMP document IDs and derivation chains between related files
Clusters	Files grouped by camera serial, camera model or author, with the field each grouping rests on
Comparison	Metadata, provenance and timing differences between two files
Explanation	Every source behind a finding for one file
Document text	The same identifiers out of document bodies, with where in the document each one was
Metadata removal	Cleaned copies of supported files, with verification of what remains
JSON output	Machine-readable results for scans and all commands
Evidence coverage	`doctor` shows which local evidence sources are available and how far back they reach

Evidence sources

Origin and activity traces

These sources are outside the analyzed file. They are records left by the operating system, browser, shell and other applications while the file was downloaded, opened, copied, extracted, synchronized or deleted.

They can reveal URLs, referrers, download times, previous paths, archive or torrent membership and other information about how the file reached or moved through the machine. Availability depends on what data still exists on the system.

Source	What `filegrail` can read
Browser download history	Download URL, referrer, time and recorded size from Chrome, Chromium, Brave, Edge, Vivaldi and Firefox profiles
Windows `Zone.Identifier`	`HostUrl`, `ReferrerUrl`, `ZoneId`
macOS Where From	Download URL and referrer stored in `kMDItemWhereFroms`
macOS quarantine	Download URL, referrer, downloading application and quarantine time
Linux XDG attributes	`user.xdg.origin.url` and `user.xdg.referrer.url`
Archives	Archive membership matched by file name and uncompressed size; archive origin can be inherited by extracted files
Torrent files and client stores	Torrent membership, trackers, client, comments, info hash/magnet data; local qBittorrent, Transmission and Deluge stores
`yt-dlp` sidecars	Page URL, uploader/channel, publication date, extractor and fetch time from `.info.json`
Shell history	Fetch commands such as `curl`, `wget`, `yt-dlp`, `scp`, `rsync`, `git`, `gh`, `aws` and others; non-fetch commands are recorded as activity
Recent documents	Linux desktop recent-file records
Windows Recent shortcuts	`.lnk` records showing that a file was opened
Sync folders	Nextcloud, Dropbox, Syncthing and OneDrive folder/account context
Trash records	Previous path and deletion time from the freedesktop trash
Messenger file names	WhatsApp and Telegram Desktop filename patterns; treated as weak evidence only

Use:

filegrail doctor

to see which of these sources are actually available on the current machine or profile.

For another user profile or a mounted copy:

filegrail doctor --home /mnt/profile
filegrail /mnt/evidence --home /mnt/profile

Supported formats

Embedded metadata

These are metadata fields stored inside the file itself. Depending on the format, filegrail can extract camera and device information, GPS coordinates, timestamps, authors and editors, creating software, document properties, media tags and authenticity metadata.

Field names are preserved as they appear in the original format, so the report shows the actual recorded metadata rather than simplified labels.

Metadata block	Extensions	Data extracted
EXIF	`.jpg` `.jpeg` `.jpe` `.tif` `.tiff` `.dng` `.nef` `.cr2` `.arw` `.orf` `.rw2` `.webp` `.heic` `.heif` `.avif`	Camera make/model, body serial, lens, software, capture time, GPS
PNG text	`.png` `.apng`	`tEXt`, `zTXt`, `iTXt`: software, creation time, author and other stored values
ISO BMFF	`.mp4` `.m4v` `.mov` `.qt` `.3gp` `.m4a` `.heic` `.heif` `.avif`	Encoder/device, creation time, ISO 6709 location
Matroska	`.mkv` `.mk3d` `.webm` `.mka`	Writing application/library, segment date, tags
RIFF/BWF	`.wav` `.wave` `.rmi` `.avi`	`LIST/INFO`, recorder information, coding history, embedded ID3 where present
Vorbis comments	`.flac` `.ogg` `.oga` `.opus` `.spx`	Vendor string and all `NAME=value` comments
ID3	`.mp3` `.aac` `.tta`	Encoding software, artist, title, date and other ID3v2 frames
PDF Info	`.pdf`	PDF `Info` dictionary
OOXML properties	`.docx` `.docm` `.dotx` `.xlsx` `.xlsm` `.xltx` `.pptx` `.pptm`	Application, author, last editor, company, template, revision count, editing time
OLE properties	`.doc` `.dot` `.xls` `.xlt` `.ppt` `.pot` `.pps` `.msg`	Summary and document-summary properties
OpenDocument metadata	`.odt` `.ods` `.odp` `.odg` `.odf` `.ott` `.otp`	Generator, author, creation and editing metadata
EPUB package	`.epub`	OPF package metadata
RTF metadata	`.rtf`	Generator and `\info` fields
SVG metadata	`.svg`	Generator and embedded RDF metadata
Jupyter notebook	`.ipynb`	Kernel name and language runtime version
C2PA	`.jpg` `.jpeg` `.png`	Producing application, creation data, digital source type and hard-binding check

Cross-format metadata

XMP, XMP history and IPTC can appear in several different file types. filegrail extracts them wherever those metadata blocks are supported, rather than treating them as belonging to one extension only.

Block	Typical data
XMP	Creating application, author, title, document IDs and derivation information
XMP history	Recorded editing steps and timestamps
IPTC	By-line, credit, source, copyright, headline, caption, keywords, place and creation date

Email

For saved email messages, filegrail can reconstruct the recorded delivery path from Received: headers, including the servers and addresses involved in each hop.

With --content, it can also inspect the readable message body in text/plain and text/html parts.

Extension	Data extracted
`.eml`	Every `Received:` hop, connecting addresses and message headers
`.msg`	Transport headers where present, plus OLE document properties

Extensions	What is read
`.zip` `.jar` `.whl`	Member names, sizes and metadata from supported files inside the archive
`.tar` `.tgz` `.gz` `.bz2` `.xz`	Members and supported metadata through the archive/compression layer

Torrents and sidecars

Torrent data can link a file to a .torrent when both the file name and exact size match an entry in the torrent.

For media downloaded with yt-dlp, a matching <name>.info.json sidecar can provide the original page URL, uploader or channel, publication date, extractor and download time.

File	Data extracted
`.torrent`	Trackers, creating client, comment, torrent membership and magnet/info-hash data
`<name>.info.json`	`yt-dlp` page URL, uploader/channel, publication date, extractor and fetch time

Document content

--content extends the scan from metadata to the actual text stored in supported documents.

The full document text is not added to the report. filegrail extracts only supported identifiers - such as URLs, domains, email addresses, IP addresses, coordinates and hashes - and records exactly where each value was found.

Content scanning is limited to 1 MB of text per file and 64 members of a packaged document.

Extensions	Content read	Place a value is reported with
`.txt` `.text` `.md` `.markdown` `.rst` `.log`	Text by line	`line 12`
`.csv` `.tsv` `.json` `.ndjson` `.jsonl` `.ipynb` `.yaml` `.yml` `.toml` `.ini` `.cfg` `.conf` `.vcf` `.ics`	Text/data by line	`line 12`
`.html` `.htm` `.xhtml` `.xml` `.svg`	Visible text and relevant URLs/attributes	`line 12`
`.docx` `.docm` `.dotx`	Body, footnotes, endnotes and comments	`body`, `footnotes`, `endnotes`, `comments`
`.xlsx` `.xlsm` `.xltx`	Shared strings and inline cell text	`cell text`, `sheet 2`
`.pptx` `.pptm`	Slide text and notes	`slide 4`, `slide 4 notes`
`.odt` `.ods` `.odp` `.odg` `.odf` `.ott` `.otp`	Document body, headers and footers	`body`, `headers and footers`
`.epub`	Chapters	the chapter's own file name
`.eml` `.msg`	Decoded message body, every text part	`body`, `body (html)`

PDF metadata is supported, but PDF body text is not extracted by --content.

Identifier types

The same identifier detection is used for metadata and, with --content, document text. Filters remove common false positives such as software versions, file names that only look like domains and hexadecimal build identifiers.

Type	Taken	Not taken
`url`	`http` and `https` addresses, normalized
`domain`	every host behind a URL or an address, and bare names whose TLD is a real one	anything shaped like a file name
`email`	addresses whose TLD is a real one	the address inside a message id - its host is still kept
`ipv4`	dotted quads, with private and reserved ranges marked as such	version numbers, and digits in a field naming software
`geo`	coordinates written with a hemisphere letter, a degree sign, a `geo:` URI, a map URL or an explicit latitude label	a bare pair of decimals, however many places it carries
`md5` `sha1` `sha256`	32, 40 and 64 hex digits	digests in a field naming software, which are build ids

Every extracted value includes the file, source and exact field or document location where it was found, so the result can be traced back to its context.

For the complete format reference and edge cases, see docs/FORMATS.md.

Analysis

Correlation and conflicts

When several sources describe the same file, filegrail compares them instead of choosing one answer automatically. Matching records can reinforce a finding; disagreements are shown as conflicts.

It can report:

multiple sources supporting the same origin
conflicting origin URLs
file-size mismatches
filename-only matches
creation/modification dates in impossible order
XMP editing steps out of sequence
EXIF vs XMP differences
PDF Info vs XMP differences
XMP derivation relationships between files
C2PA hard-binding mismatches

How a record was matched to a file

Every record says how it came to be about the file it is reported under. A download row found by the path it was saved to and one found by a name that happens to be the same are not equally firm, and the report never leaves that to be guessed.

Basis	What it means	Where it comes from
`read from the file`	Decoded out of the file's own bytes	EXIF, XMP, IPTC, document properties, Content Credentials, mail headers
`file attribute`	Read from what the filesystem keeps for this exact file	`Zone.Identifier`, macOS where-from, XDG attributes, creation times
`recorded path`	An external store names this exact path	Browser download history, Recent Documents
`record beside the file`	A separate file written next to it and naming it	`yt-dlp` sidecar, freedesktop trash record
`name and exact size`	Both agree; two files can share both, but not easily	Torrents, archive members, Windows shortcuts
`file name`	The name is all that matched	A download record for a file that has since moved, a messenger naming pattern
`container member`	Read from a member, or inherited from the container	Archives
`inside a sync folder`	The file lies under a folder a client manages	Nextcloud, Dropbox, Syncthing, OneDrive

filegrail does not put a number on how much a record is worth. There is no probability behind such a number and no forensic basis for one; what it reports instead is the category, the source and the basis of the match, which are facts.

Extracting investigation pivots from metadata and content

filegrail ./case --identify      # out of metadata
filegrail ./case --content       # out of metadata and document text

--identify extracts the supported identifier types from metadata and provenance records.

--content does the same for the body of supported documents and automatically enables identifier extraction.

Values are grouped by type - urls, domains, emails, ip addresses, coordinates, hashes - one row per place a value was seen, with the file, the source and the field or location it came from. A value seen under more than one source gets a section of its own at the end: an address written in a document and recorded in how the file arrived is the pairing reading content exists to find.

Clusters

filegrail ./photos --cluster

Finds files that share an identifying value, and says which field the grouping rests on:

camera serial - EXIF · BodySerialNumber. A serial is assigned per unit, so the same one is the same physical camera.
camera model - EXIF · Make + Model. A product line thousands of people own, which is not the same claim.
author - OOXML · creator or the equivalent. A name somebody typed.

Timeline

filegrail ./case --timeline

Combines available timestamps from origin records, embedded metadata and later activity into one chronological view.

File relationships

XMP identifiers such as:

xmpMM:DocumentID
xmpMM:InstanceID
xmpMM:OriginalDocumentID
xmpMM:DerivedFrom

can reveal relationships between files even after they have been renamed or exported. The scan can identify relationships such as derived-from, source-of, same-document and common-ancestor.

Installation

Requires Python 3.10+.

pipx install filegrail

or:

uv tool install filegrail

From the repository:

git clone https://github.com/osint-shifu/filegrail.git
cd filegrail
PYTHONPATH=src python -m filegrail.cli /path/to/files

Runtime dependencies: 0.

Usage

filegrail <path> [options]
filegrail <command> [options]

Analyze one file:

filegrail suspicious.pdf

Analyze a directory recursively:

filegrail ./evidence

Analyze the current directory:

filegrail .

Running filegrail with no arguments displays the command overview and does not start a scan.

Commands

filegrail PATH is the normal scan command. filegrail scan PATH does the same thing explicitly.

Use explain when you want the evidence behind one file, compare for two files and doctor to check which provenance sources are available on the machine.

Command	Purpose
`filegrail PATH`	Scan one file or directory
`filegrail scan PATH`	Explicit scan command
`filegrail explain FILE`	Show the evidence behind findings for one file
`filegrail compare A B`	Compare two files
`filegrail doctor`	Show available local evidence sources
`filegrail clean PATH --out DIR`	Write metadata-cleaned copies
`filegrail clean PATH --check`	Check what cleaning would remove without writing files
`filegrail menu`	Interactive command menu
`filegrail help COMMAND`	Command-specific help

Scan options

A normal scan checks embedded metadata and available local provenance traces. Additional work such as document-content inspection, SHA-256 hashing and clustering is enabled only when requested.

--home lets the same analysis use another user profile, copied profile or mounted system image as the source of local evidence.

Option	Purpose
`-v`, `--verbose`	Show every evidence record
`--brief`	Index only, without per-file detail
`--timeline`	Chronological event view
`--identify`	Extract investigation identifiers
`--content`	Also inspect supported document content; implies `--identify`
`--cluster`	Group files by shared cameras/authors
`--unknown-only`	Show only files with no evidence found
`--hash`	Compute SHA-256 for each file
`-j`, `--json`	JSON output
`--redact`	Redact credentials before printing
`--type NAME`	Filter by `archive`, `audio`, `document`, `image`, `mail`, `text` or `video`
`--ext LIST`	Filter by extensions, e.g. `--ext jpg,pdf`
`--limit N`	Limit files with no evidence found; `0` means all
`--home DIR`	Read evidence from another user profile
`--no-recurse`	Do not scan subdirectories
`--no-skip`	Include normally skipped build/cache/vendor directories
`--no-shell-history`	Disable shell-history correlation
`--no-archives`	Disable archive-origin inheritance
`--color`	Force ANSI color
`--no-color`	Disable ANSI color

Clean options

Use --out DIR to write cleaned copies to another directory. Use --check to see what would be removed and whether anything would remain, without writing files.

The output directory cannot be inside the directory being cleaned.

Option	Purpose
`--out DIR`	Output directory for cleaned copies
`--check`	Check cleaning without writing files
`--overwrite`	Replace an existing destination file
`--type NAME`	Filter by file family
`--ext LIST`	Filter by extension
`--no-recurse`	Do not descend into subdirectories
`-j`, `--json`	JSON output

Examples

# Full analysis of one file
filegrail photo.jpg

# Scan a case and show only the index
filegrail ./case --brief

# Extract investigation pivots
filegrail ./case --identify

# Also read the text inside supported documents
filegrail ./case --content

# Build a timeline
filegrail ./case --timeline

# Find files sharing cameras or authors
filegrail ./case --cluster

# Explain exactly why a finding was produced
filegrail explain document.pdf

# Compare two files
filegrail compare original.docx edited.docx

# Analyze another user profile
filegrail /mnt/evidence --home /mnt/profile

# JSON report
filegrail ./case --json > report.json

# JSON report with credentials redacted
filegrail ./case --redact --json > report.json

Example views

Below are examples of actual terminal output for the main workflows: scanning one file, scanning a directory, resolving conflicts, explaining evidence, building a timeline, extracting identifiers, clustering files, comparing files and checking evidence coverage.

One file · filegrail holiday.jpg

filegrail 0.8.0
────────────────────────────────────────────────────────────────────────
target    ~/case/press/holiday.jpg · profile ~/home · external

FILE  ·  holiday.jpg · JPEG · 3.4 MB
────────────────────────────────────────────────────────────────────────

  path    /home/oryon/case/press/holiday.jpg
  mtime   2026-09-06 00:40:40

ORIGIN  ·  1 record
────────────────────────────────────────────────────────────────────────

  source             match          time
  ─────────────────  ─────────────  ───────────────────
› Chromium download  recorded-path  2026-08-31 10:49:33
  │ url       https://portal.example.org/press/2026/holiday-master.jpg
  └ referrer  https://portal.example.org/press/

METADATA  ·  1 source · 8 fields
────────────────────────────────────────────────────────────────────────

  source  summary
  ──────  ──────────────────────────────────────────────────────────────
› EXIF    NIKON COOLPIX P6000 · serial 3001234 · 43.467447, 11.885128 ·
          2008-10-22 16:28:39

EXIF  ·  8 fields
────────────────────────────────────────────────────────────────────────

  field             value
  ────────────────  ───────────────────
  Make              NIKON
  Model             COOLPIX P6000
  DateTimeOriginal  2008:10:22 16:28:39
  BodySerialNumber  3001234
  GPSLatitudeRef    N
  GPSLatitude       43, 28, 2.81
  GPSLongitudeRef   E
  GPSLongitude      11, 53, 6.46

This view separates how the file arrived from what the file says about itself. The meter shows how directly each source supports the finding.

A whole directory, top to bottom · filegrail ./case

filegrail 0.8.0
────────────────────────────────────────────────────────────────────────
target    ~/case · profile ~/home · external

SUMMARY  ·  4 files · 4 types · 3.4 MB
────────────────────────────────────────────────────────────────────────

  with evidence      3
  unresolved         1
  origin records     2
  metadata sources   3
  activity records   0
  findings           0

FILES  ·  4 files · 4 types · 3.4 MB
────────────────────────────────────────────────────────────────────────

  file               type  size    origin             metadata
  ─────────────────  ────  ──────  ─────────────────  ────────────────
  press/holiday.jpg  JPEG  3.4 MB  Chromium download  EXIF
  invoice.docx       DOCX  834 B   XDG attribute      OOXML properties
  chart.png          PNG   88 B    —                  PNG text
· notes.md           MD    64 B    —                  —

  · no evidence found

ORIGIN  ·  2 records · 2 files
────────────────────────────────────────────────────────────────────────

  file            source             match           time
  ──────────────  ─────────────────  ──────────────  ───────────────────
› press/holiday.  Chromium download  recorded-path   2026-08-31 10:49:33
  jpg
  │ url       https://portal.example.org/press/2026/holiday-master.jpg
  └ referrer  https://portal.example.org/press/
› invoice.docx    XDG attribute      file-attribute  —
  └ url  https://acme-legal.example/portal/invoice.docx

METADATA  ·  3 sources · 3 files
────────────────────────────────────────────────────────────────────────

  file               source            summary
  ─────────────────  ────────────────  ─────────────────────────────────
› press/holiday.jpg  EXIF              NIKON COOLPIX P6000 · serial
                                       3001234 · 43.467447, 11.885128 ·
                                       2008-10-22 16:28:39
› invoice.docx       OOXML properties  Ann Shaw
› chart.png          PNG text          GIMP 2.10

UNRESOLVED  ·  1 file
────────────────────────────────────────────────────────────────────────

  file          type  size  last modified
  ────────────  ────  ────  ───────────────────
· notes.md      MD    64 B  2026-09-06 00:40:40

SCAN GAPS  ·  1 item
────────────────────────────────────────────────────────────────────────

  what             detail
  ───────────────  ───────────────────────────────────
· browser history  2 download records across 1 profile

A directory scan starts with a summary and file index, then shows detailed findings for each file and the evidence sources that produced them.

Index only, for a large directory · filegrail ./case --brief

filegrail 0.8.0 · brief
────────────────────────────────────────────────────────────────────────
target    ~/case · profile ~/home · external

SUMMARY  ·  4 files · 4 types · 3.4 MB
────────────────────────────────────────────────────────────────────────

  with evidence   3
  unresolved      1
  needs review    0

FILES  ·  4 files · 4 types · 3.4 MB
────────────────────────────────────────────────────────────────────────

  file               type  size    origin             metadata
  ─────────────────  ────  ──────  ─────────────────  ────────────────
  press/holiday.jpg  JPEG  3.4 MB  Chromium download  EXIF
  invoice.docx       DOCX  834 B   XDG attribute      OOXML properties
  chart.png          PNG   88 B    —                  PNG text
· notes.md           MD    64 B    —                  —

  · no evidence found

--brief keeps the scan to one row per file. Files with no provenance or metadata findings are still listed.

Two records that disagree · filegrail ./contested

filegrail 0.8.0
────────────────────────────────────────────────────────────────────────
target    ~/contested/statement.pdf · profile ~/home · external

FILE  ·  statement.pdf · PDF · 65 B
────────────────────────────────────────────────────────────────────────

  path    /home/oryon/contested/statement.pdf
  mtime   2026-09-06 00:40:40

ORIGIN  ·  2 records
────────────────────────────────────────────────────────────────────────

  source             match           time
  ─────────────────  ──────────────  ───────────────────
› Chromium download  recorded-path   2026-08-31 10:49:33
  │ url       https://documents.example.org/releases/statement.pdf
  └ referrer  https://documents.example.org/releases/
› XDG attribute      file-attribute  —
  └ url  https://mail.example.net/attach/statement.pdf

METADATA  ·  1 source · 1 field
────────────────────────────────────────────────────────────────────────

  source    summary
  ────────  ────────────────
› PDF Info  LibreOffice 24.2

FINDINGS  ·  2 findings · 2 needs review
────────────────────────────────────────────────────────────────────────

  type             file           field       sources
  ───────────────  ─────────────  ──────────  ───────
! source conflict  statement.pdf  origin URL  —
  └   browser download says
      https://documents.example.org/releases/statement.pdf
! source conflict  statement.pdf  origin URL  —
  └   XDG attribute says https://mail.example.net/attach/statement.pdf

  ! needs review

PDF INFO  ·  1 field
────────────────────────────────────────────────────────────────────────

  field         value
  ────────────  ────────────────
  Producer      LibreOffice 24.2

When two independent records disagree, both remain visible and the report marks the conflict instead of choosing one automatically.

Why a finding was produced · filegrail explain holiday.jpg

filegrail 0.8.0 · explain
────────────────────────────────────────────────────────────────────────
target    ~/case/press/holiday.jpg · profile ~/home · external

SUMMARY  ·  2 evidence records
────────────────────────────────────────────────────────────────────────

  origin records     1
  metadata sources   1
  activity records   0
  correlation        0

ORIGIN  ·  1 record
────────────────────────────────────────────────────────────────────────

  source             match          time
  ─────────────────  ─────────────  ───────────────────
› Chromium download  recorded-path  2026-08-31 10:49:33
  │ url       https://portal.example.org/press/2026/holiday-master.jpg
  └ referrer  https://portal.example.org/press/

METADATA  ·  1 source · 8 fields
────────────────────────────────────────────────────────────────────────

  source  summary
  ──────  ──────────────────────────────────────────────────────────────
› EXIF    NIKON COOLPIX P6000 · serial 3001234 · 43.467447, 11.885128 ·
          2008-10-22 16:28:39

EXIF  ·  8 fields
────────────────────────────────────────────────────────────────────────

  field             value
  ────────────────  ───────────────────
  Make              NIKON
  Model             COOLPIX P6000
  DateTimeOriginal  2008:10:22 16:28:39
  BodySerialNumber  3001234
  GPSLatitudeRef    N
  GPSLatitude       43, 28, 2.81
  GPSLongitudeRef   E
  GPSLongitude      11, 53, 6.46

explain shows the assessment first, followed by the origin, metadata and activity records behind it.

Chronological events · filegrail ./case --timeline

filegrail 0.8.0 · timeline
────────────────────────────────────────────────────────────────────────
target    ~/case · profile ~/home · external

TIMELINE  ·  2 events · 1 file
────────────────────────────────────────────────────────────────────────

  time                 file               source             event
  ───────────────────  ─────────────────  ─────────────────  ──────────
  2008-10-22 16:28:39  press/holiday.jpg  EXIF               captured
  2026-08-31 10:49:33  press/holiday.jpg  Chromium download  downloaded

--timeline places all available dated events on one chronological axis, whether they came from an origin record, from metadata or from later activity.

Identifiers, including document content · filegrail ./case --content

IDENTIFIERS  ·  9 unique values · 11 occurrences · 1 cross-source
────────────────────────────────────────────────────────────────────────

URLS  ·  3 unique values · 3 occurrences
────────────────────────────────────────────────────────────────────────

› https://acme-legal.example/portal/invoice.docx
  │ file    invoice.docx
  │ source  XDG attribute
  └ where   url
› https://portal.example.org/press
  │ file    holiday.jpg
  │ source  browser download
  └ where   referrer
› https://portal.example.org/press/2026/holiday-master.jpg
  │ file    holiday.jpg
  │ source  browser download
  └ where   url

DOMAINS  ·  3 unique values · 5 occurrences
────────────────────────────────────────────────────────────────────────

  value               file          source            field / location
  ──────────────────  ────────────  ────────────────  ────────────────
  acme-legal.example  invoice.docx  XDG attribute     url
  acme-legal.example  invoice.docx  content           body
  portal.example.org  holiday.jpg   browser download  url
  portal.example.org  holiday.jpg   browser download  referrer
  innafirma.example   notes.md      content           line 1

EMAILS  ·  2 unique values · 2 occurrences
────────────────────────────────────────────────────────────────────────

  value                        file          source   field / location
  ───────────────────────────  ────────────  ───────  ────────────────
  ann.shaw@acme-legal.example  invoice.docx  content  body
  kontakt@innafirma.example    notes.md      content  line 1

COORDINATES  ·  1 unique value · 1 occurrence
────────────────────────────────────────────────────────────────────────

  value              file         source           field / location
  ─────────────────  ───────────  ───────────────  ────────────────
  43.46745,11.88513  holiday.jpg  device metadata  geo

CROSS-SOURCE MATCHES  ·  1 value
────────────────────────────────────────────────────────────────────────

  value               sources                  files
  ──────────────────  ───────────────────────  ────────────
  acme-legal.example  XDG attribute · content  invoice.docx

UNRESOLVED  ·  1 file
────────────────────────────────────────────────────────────────────────

  file          type  size  last modified
  ────────────  ────  ────  ───────────────────
· notes.md      MD    64 B  2026-09-06 00:40:40

SCAN GAPS  ·  1 item
────────────────────────────────────────────────────────────────────────

  what             detail
  ───────────────  ───────────────────────────────────
· browser history  2 download records across 1 profile

Identifier results show their origin: recorded for metadata or provenance, text for document content and both when the same value appears in both.

Files sharing a camera or an author · filegrail ./shots --cluster

CLUSTERS  ·  2 groups · 3 files
────────────────────────────────────────────────────────────────────────

  attribute      value                files  basis
  ─────────────  ───────────────────  ─────  ───────────────────────
› camera serial  3001234                  3  EXIF · BodySerialNumber
  │ dune-01.jpg
  │ dune-02.jpg
  └ harbour.jpg
› camera model   NIKON COOLPIX P6000      3  EXIF · Make + Model
  │ dune-01.jpg
  │ dune-02.jpg
  └ harbour.jpg

A shared camera serial can point to the same physical device. A shared camera model only shows that the files name the same model.

Two files against each other · filegrail compare beach.jpg beach-edited.jpg

filegrail 0.8.0 · compare
────────────────────────────────────────────────────────────────────────
left      /home/oryon/beach.jpg
right     /home/oryon/beach-edited.jpg

FILES  ·  2 files
────────────────────────────────────────────────────────────────────────

  field         beach.jpg  beach-edited.jpg
  ────────────  ─────────  ────────────────
  type          JPEG       JPEG
  size          335 B      368 B

METADATA  ·  4 compared fields
────────────────────────────────────────────────────────────────────────

  field             beach.jpg            beach-edited.jpg
  ────────────────  ───────────────────  ────────────────────
  Make              NIKON                NIKON
  Model             COOLPIX P6000        COOLPIX P6000
  BodySerialNumber  3001234              3001234
  Software          NIKON COOLPIX P6000  Adobe Photoshop 26.1

ORIGIN  ·  2 records
────────────────────────────────────────────────────────────────────────

  file              source
  ────────────────  ────────────────
  beach.jpg         no origin record
  beach-edited.jpg  no origin record

CORRELATION  ·  5 results
────────────────────────────────────────────────────────────────────────

  result      field             value
  ──────────  ────────────────  ────────────────────────────────────────
  match       Make              NIKON
  match       Model             COOLPIX P6000
  match       BodySerialNumber  3001234
! difference  Software          NIKON COOLPIX P6000 ≠ Adobe Photoshop
                                26.1
  interval    created           0 seconds apart

RELATIONSHIPS  ·  1 relation
────────────────────────────────────────────────────────────────────────

  relationship  files                         basis
  ────────────  ────────────────────────────  ────────────────────────
  same device   beach.jpg · beach-edited.jpg  BodySerialNumber 3001234

What this machine can answer at all · filegrail doctor

filegrail 0.8.0 · doctor
────────────────────────────────────────────────────────────────────────
profile   ~/home · external

SUMMARY  ·  13 sources
────────────────────────────────────────────────────────────────────────

  available     4
  partial       1
  unavailable   8

SOURCES  ·  13 sources · 4 available · 1 partial · 8 unavailable
────────────────────────────────────────────────────────────────────────

  source                     type        status       coverage / detail
  ─────────────────────────  ──────────  ───────────  ──────────────────
  Chromium family downloads  artifact    available    2 records across 1
                                                      of 1 profile
  Firefox downloads          artifact    unavailable  no profile found
  XDG origin attribute       artifact    available    written by KDE
                                                      tools and wget
                                                      --xattr, but not
                                                      by Firefox
  Mounted Zone.Identifier    artifact    available    user.Zone.Identifi
                                                      er on an NTFS
                                                      mount
  Shell history              artifact    unavailable  no history file
                                                      found
  Recent documents           artifact    unavailable  no list found
  macOS quarantine database  artifact    unavailable  no database in
                                                      this profile
  Deleted files              artifact    unavailable  no trash directory
  Windows Recent shortcuts   artifact    unavailable  no Recent folder
                                                      in this profile
  Torrent client stores      artifact    unavailable  no client store
                                                      found
  Sync client folders        artifact    unavailable  no client
                                                      configuration
                                                      found
  Creation timestamps        filesystem  available    statx
  C2PA signature check       parser      partial      manifests are read
                                                      and their hash
                                                      binding
                                                      recomputed;
                                                      validating the
                                                      certificate chain
                                                      needs a crypto
                                                      library

LIMITATIONS  ·  2 items
────────────────────────────────────────────────────────────────────────

  source                  limitation
  ──────────────────────  ──────────────────────────────────────────────
  Chromium family record  no record before 2026-08-31
  C2PA signature check    manifests are read and their hash binding
                          recomputed; validating the certificate chain
                          needs a crypto library

doctor distinguishes between a source that was available but contained no matching record and a source that was not available to search at all.

Checking before publishing · filegrail clean ./case --check

filegrail 0.8.0 · clean --check
────────────────────────────────────────────────────────────────────────
target    ~/case
mode      check only · nothing written

SUMMARY  ·  4 files
────────────────────────────────────────────────────────────────────────

  cleanable      3
  unsupported    1
  would remain   0
  exit           0

RESULTS  ·  4 files · 3 cleanable · 1 unsupported
────────────────────────────────────────────────────────────────────────

  file               format  metadata             result
  ─────────────────  ──────  ───────────────────  ───────────
  chart.png          PNG     PNG text             would clean
  invoice.docx       DOCX    document properties  would clean
  notes.md           MD      —                    unsupported
  press/holiday.jpg  JPEG    EXIF                 would clean

Exit code 0 if every copy would come out clean, 1 if any would not.

Metadata removal

filegrail clean removes supported metadata from copies. Originals are never modified.

Cleanable formats

Only the formats listed below can be cleaned. Unsupported files are left unchanged and reported as such.

Cleaned files are written as separate copies. The originals are never modified.

Family	Extensions
JPEG	`.jpg` `.jpeg` `.jpe`
PNG	`.png` `.apng`
ISO BMFF media	`.mp4` `.m4v` `.m4a` `.mov` `.qt` `.3gp`
Microsoft OOXML	`.docx` `.docm` `.dotx` `.xlsx` `.xlsm` `.xltx` `.pptx` `.pptm`
OpenDocument	`.odt` `.ods` `.odp` `.odg` `.ott` `.otp`

Clean one file:

filegrail clean photo.jpg --out ./clean

Check without writing:

filegrail clean ./publish --check

After cleaning, filegrail scans each copy again. If supported metadata is still detectable, it is reported instead of being treated as successfully removed.

Metadata removal is not anonymization. Image pixels, sensor patterns, codec fingerprints and document content can still identify a source.

JSON and automation

Use --json when results need to be processed by scripts, jq, notebooks, pipelines or other tools. All main commands support machine-readable output with command-specific schemas and exit codes.

Schemas:

filegrail.scan/1
filegrail.explain/1
filegrail.compare/1
filegrail.doctor/1
filegrail.clean/1

What a script gets back:

Code	Meaning
`0`	The command ran. For `clean` and `clean --check`, every copy came out clean
`1`	`clean` only: metadata survived in at least one copy, or would
`2`	The command was asked for something it cannot do - a missing path, two arguments where one file was needed, an unknown option

JSON output preserves:

files and paths
origin, metadata and activity records
decoded metadata
evidence sources
the match basis behind every record
findings and conflicts
extracted identifiers
clusters of files sharing an attribute
file relationships

Privacy

filegrail performs analysis locally and makes no network requests. The main privacy risk is therefore the report itself, which may contain sensitive information recovered from local files and system traces.

Reports can contain sensitive local data such as:

private URLs
credentials or tokens embedded in URLs/commands
file-system paths
email addresses
IP addresses
GPS coordinates

Use:

filegrail ./case --redact

or:

filegrail ./case --redact --json

before sharing output.

Read a redacted report yourself before you publish it.

Limits

filegrail can only use evidence that still exists.
Cleared browser history, removed extended attributes or missing shell history cannot be reconstructed.
A supported sync folder can show account/folder context, but not who uploaded a file.
WhatsApp/Telegram filename patterns do not identify a sender or conversation.
C2PA hard binding is checked, but the certificate chain/signature trust is not verified.
PDF metadata is read; PDF body text is not extracted by --content.
Unsupported file formats can still participate in provenance analysis when external/local evidence about them exists.
filegrail is not a monitoring agent, chain-of-custody system or full disk-forensics suite.

Documentation

License

Apache-2.0.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

filegrail-0.8.0.tar.gz (403.1 kB view details)

Uploaded Sep 6, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

filegrail-0.8.0-py3-none-any.whl (224.3 kB view details)

Uploaded Sep 6, 2026 Python 3

File details

Details for the file filegrail-0.8.0.tar.gz.

File metadata

Download URL: filegrail-0.8.0.tar.gz
Upload date: Sep 6, 2026
Size: 403.1 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for filegrail-0.8.0.tar.gz
Algorithm	Hash digest
SHA256	`d657c6e85cda35e5e9c6778a1e4074d4cea346c517d187f3765a527dfd6c6531`
MD5	`42a987572ee8c2d32080f3dc12d8532e`
BLAKE2b-256	`e694a2bcb26d5ef7ff40463bd27f8e38d476f129cd9ea69967c417164a381d9b`

See more details on using hashes here.

Provenance

The following attestation bundles were made for filegrail-0.8.0.tar.gz:

Publisher: release.yml on osint-shifu/filegrail

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: filegrail-0.8.0.tar.gz
- Subject digest: d657c6e85cda35e5e9c6778a1e4074d4cea346c517d187f3765a527dfd6c6531
- Sigstore transparency entry: 2732831310
- Sigstore integration time: Sep 6, 2026
Source repository:
- Permalink: osint-shifu/filegrail@bd920682c4bcdd0b200adea126b7013b596bf4ee
- Branch / Tag: refs/tags/v0.8.0
- Owner: https://github.com/osint-shifu
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: release.yml@bd920682c4bcdd0b200adea126b7013b596bf4ee
- Trigger Event: push