Backed by mandō

Know every file.
Instantly.

MAGIQ is the universal file intelligence platform. Powered by magic number analysis, MIME detection, and deep content classification — we tell you exactly what any file is. No extensions. No guessing. Just truth.

Try the Scanner
M K S Cole Dupont Tyler Nakamura
6,800+ teams identifying files at scale
MAGIQ Scanner
📄
document.pdf
TYPE PDF document, version 1.7 application/pdf

Trusted by teams at

⚡ Vercel
🔷 Docker
🟠 AWS S3
🐙 GitHub
🔶 Cloudflare
🟣 Datadog
🔵 Kubernetes
🟢 Slack

Features

Every file has a story.
We read it in milliseconds.

From magic number analysis to MIME classification, MAGIQ gives you total file intelligence with a single command.

🔮

MagicSense™ core

Our engine reads magic numbers — the invariant byte signatures hidden at fixed offsets in every file. We match against 20,000+ patterns in the world's most comprehensive magic database. No file extension required.

🏷️

MIMEVision™ -i

Output proper MIME type strings instead of human-readable descriptions. Get application/pdf instead of "PDF document." Standards-compliant content typing for your automated pipelines.

🔤

EncodeSense™ --mime-encoding

Detect the character encoding of any file — US-ASCII, UTF-8, UTF-16, ISO-8859, EBCDIC. Stop guessing why your text renders wrong. EncodeSense™ delivers encoding truth in one pass.

📦

DeepScan™ -z

Look inside compressed files without decompressing them first. Gzip, bzip2, xz — our engine peers through the compression layer to identify what's actually inside. X-ray vision for archives.

🔗

LinkTrace™ -L

Follow symbolic links to identify the actual target file, not just the symlink itself. Because when you ask "what is this file?" you mean the destination, not the shortcut.

⚡

SpecialOps™ -s

Read and classify block and character special files — raw disk partitions, device nodes, the files that ordinary tools refuse to touch. Determine filesystem types from raw partitions.

📋

BatchProbe™ -f

Read file names from a list and classify them all in a single invocation. Process thousands of files from a manifest. Enterprise-grade batch identification with zero scripting overhead.

🎯

MultiMatch™ -k

Don't stop at the first match. Keep going and report every possible type classification, ranked by confidence. When ambiguity exists, MultiMatch™ shows you all the candidates.

🏗️

MagicForge™ -C

Compile custom magic files into optimized binary format for lightning-fast lookups. Define your own file signatures, build your magic database, and deploy it to your fleet.

🧹

InstantID™ -b

Brief mode. No filenames prepended. Just the raw type identification — clean, machine-parseable output ready to pipe into your next tool. Maximum signal, minimum ceremony.

🔬

FilterEngine™ -e

Exclude specific tests from the detection pipeline — apptype, ascii, encoding, elf, tar, soft magic. Tune exactly which analyses run for precision results and optimal performance.

📚

CustomMagic™ -m

Specify alternate magic files or directories. Override the system database with your own proprietary patterns. Because your organization's files deserve organization-specific intelligence.

Manifesto

File extensions are a lie.

Every day, billions of files move through systems with nothing but a three-letter extension to declare their identity. Rename a .exe to .jpg. Change a .pdf to .txt. The world keeps spinning. And your security stack? It has no idea.

We started MAGIQ because we believe file identity should be intrinsic, not cosmetic. The magic number — those invariant bytes hidden at fixed offsets — is the only source of truth. It's been there since Research Unix V4 in 1973. Fifty years of ignored wisdom.

MAGIQ performs three tests in sequence: filesystem tests (stat the file), magic tests (read the binary signature), and language tests (analyze text content). The first test that succeeds determines the type. No guessing. No heuristics. Just deterministic classification backed by the world's most comprehensive magic database.

We're not building a file extension lookup table. We're building universal file intelligence. And every file on earth is about to get a lot less mysterious.

How It Works

Three tests. One truth.

01

Filesystem Analysis

MAGIQ starts by examining the stat(2) return. Is it empty? A socket? A symlink? A named pipe? A block device? Filesystem-level classification catches the obvious before we go deeper.

02

Magic Number Matching

The core engine reads bytes from fixed offsets and matches them against 20,000+ signatures in our compiled magic database. ELF binaries, PDF headers, JPEG SOI markers — every known magic number, identified in microseconds.

03

Language & Encoding Detection

For text files, MAGIQ detects the character encoding (ASCII, UTF-8, UTF-16, EBCDIC) and analyzes content for language patterns — C keywords, troff macros, shell script indicators.

04

Instant Classification

The first successful test produces the type. Output as human-readable text, MIME type, or both. Pipe it, parse it, or feed it to your automation. Universal file intelligence, delivered.

File Scanner

Drop any file.
Watch MAGIQ work.

Our drag-and-drop scanner identifies file types in real time. See the magic number analysis happen before your eyes.

🔮

Drop files here to identify

or click to simulate

MIME Type Registry

The complete MIME
classification index.

Browse every MIME type MAGIQ recognizes. Search by type, extension, or magic signature.

Magic Database

Explore 20,000+ magic patterns.

The beating heart of MAGIQ — our compiled magic database maps byte signatures to file types.

/usr/share/misc/magic
20,847
Magic patterns
1,200+
File types
450+
MIME types
v5.46
Database version
0
Files identified daily
0
Classification accuracy
0
Average scan time
0
Magic patterns loaded

Demo

See MAGIQ in action.

Real commands. Real output. Real file intelligence.

~/projects

Pricing

Universal intelligence.
Universally accessible.

Starter
$0/mo

For individuals exploring file intelligence

  • Single file identification
  • Basic MagicSense™
  • InstantID™ brief mode
  • Community magic database
  • 500 scans/day
Most Popular
Pro
$39/mo

For teams that need deep file intelligence

  • Unlimited identification
  • MIMEVision™ + EncodeSense™
  • DeepScan™ compressed files
  • LinkTrace™ + SpecialOps™
  • BatchProbe™ list processing
  • Priority support
Enterprise
$249/mo

For organizations at classification scale

  • Everything in Pro
  • MultiMatch™ full analysis
  • MagicForge™ custom databases
  • FilterEngine™ test control
  • CustomMagic™ patterns
  • SSO & RBAC
  • Dedicated CSM
On-Prem
Custom

For sovereign file intelligence

  • Everything in Enterprise
  • Self-hosted deployment
  • Private magic databases
  • Air-gapped operation
  • Compliance certifications
  • White-glove onboarding

Testimonials

Teams that know their files
don't get surprised.

★★★★★

"Someone renamed a shell script to .png and uploaded it to our asset pipeline. MAGIQ caught it instantly. MagicSense™ doesn't care about extensions — it reads the actual bytes. That one catch prevented a supply chain attack."

Sylvie Whitfield
Sylvie Whitfield Head of Security, ShieldOps
★★★★★

"We process 2M uploads per day. Before MAGIQ, we relied on file extensions and Content-Type headers — both trivially spoofable. MIMEVision™ gives us ground-truth MIME types from the actual file content. Game changer."

Amara Okonkwo
Amara Okonkwo VP Engineering, CloudVault
★★★★★

"DeepScan™ is witchcraft. We had a .tar.gz that was actually a bzip2-compressed ISO image someone misnamed three jobs ago. MAGIQ looked through the compression and told us exactly what was inside. Saved us 4 hours of debugging."

Felix Chen
Felix Chen Principal SRE, InfraPrime
★★★★★

"We built a custom magic database with MagicForge™ to identify our proprietary binary format. Compiled it, deployed it to 400 nodes, and now every file in our pipeline gets classified with our internal taxonomy. The ROI is absurd."

Cole Dupont
Cole Dupont Platform Lead, DataForge
★★★★★

"EncodeSense™ ended the encoding wars on our team. No more guessing if a CSV is UTF-8 or ISO-8859-1. We pipe every inbound file through MAGIQ first, and the encoding detection alone justified the Enterprise tier."

Tyler Nakamura
Tyler Nakamura CTO, ParseKit

FAQ

Questions about magic,
answered without magic.

A magic number is a fixed sequence of bytes at a specific offset in a file that uniquely identifies its type. For example, PDF files always start with %PDF, PNG files with \x89PNG, and ELF binaries with \x7fELF. MAGIQ maintains a database of 20,000+ such patterns — it's how we identify files without relying on extensions.

File extensions are metadata — cosmetic labels that anyone can change. MAGIQ reads the actual content of the file, examining magic numbers, byte patterns, and text encoding. A .jpg that's actually an executable will be correctly identified as an executable. Extensions lie. Magic numbers don't.

Yes. DeepScan™ (-z) peers inside compressed files to identify their actual contents. A gzipped tar archive will be reported as a tar archive, not just "gzip compressed data." We support gzip, bzip2, xz, and more compression formats.

Absolutely. With CustomMagic™ (-m), specify alternate magic files. With MagicForge™ (-C), compile your custom patterns into optimized binary format. Organizations with proprietary file formats use this to extend MAGIQ's intelligence to their own ecosystem.

Never. MAGIQ is 100% read-only. With -p (preserve-date), we even restore the access time, so it looks like we never touched the file at all. Pure observation. Zero mutations. Your files remain exactly as they were.

--mime-type gives you the content type (e.g., text/plain, application/pdf). --mime-encoding gives you the character encoding (e.g., us-ascii, utf-8). Use -i to get both combined as text/plain; charset=us-ascii. Different questions, precise answers.

Every file has an identity.
It's time you knew it.

Join 6,800+ teams who've already secured early access to MAGIQ. Founding members get lifetime Pro pricing and custom magic pattern support.

No spam. No misidentified emails. Just file intelligence updates.

See mandō's portfolio

The accelerator behind the tools that run the internet.

Lyst Drift Locus Repliq Portage Voidal Scaffold Vøid Imprint Concat Stratum Folio Topline Tailstream Sieve Seekr Rewrite Trellis Sortable SINGULR Census Slice Morph FanOut Argonaut Reverb Formatik Epoch Solstice IDENT Signum Quorum Pulsar Overseer Terminus PURGE Dispatch Sotto Limelight Persist Vantage Canopy Attaché Détaché Tether Aegis CLAIMR Cohort Archivex Deflate Unravel ArcVault XTRAKT Conduit Fetch VaultLink Teleport DeltaSync Heartbeat Hopscotch Resolver Namesake Uplink CONNEX Callsign Milieu Propagate Sourcery Persona SEVER Chronicle Codex Relevance Resolv Nomenclature AFFIRMATIV file Intrinsic Diverge Patchwork VIGIL Breeze Cadence Kronos Atelier SUPRA Mandate Turnkey Accession Reforge Collective PipeForge Netweave Panoptik Substrate Repose Verity FAULTR Provenance Chrono