AI FORMAT ANALYSIS

AI Detect Base64, JSON, JWT and File Formats

Identify pasted Base64, JSON, JWT, XML, HTML, CSS, image, PDF, byte-array and encoded text, then open the most relevant converter or formatter.

Base64 / Data URLJSON / JWT / XMLPNG / WebP / AVIF / HEICPDF / Office / archivesBytes / hex
Input

AI Smart Detector

Detect Base64, JSON, JWT, Image and File Formats

The AI Format Detector helps identify pasted content before you choose a converter. It checks text structure, Data URI metadata, Base64-decoded signatures, byte arrays, hexadecimal bytes and common web file formats, then suggests the most relevant FileFlow Tool converter, formatter or analyzer.

How to use this tool

  1. Paste your text into the input field.
  2. Click the action button to process it.
  3. Copy or download the result.

Privacy

Processed locally; nothing is uploaded.

Detect unknown pasted data before converting

Use this page when you are not sure whether copied content is Base64, a Data URI, byte array, hexadecimal bytes, JSON, JWT, XML, SVG, HTML, CSS, a URL or URL-encoded text.

The detector also inspects decoded file signatures for PNG, JPEG, GIF, WebP, BMP, TIFF, ICO, AVIF, HEIC/HEIF still images and sequences, PDF, ZIP, DOCX, XLSX, PPTX, fonts and common media containers.

Recommendations matched to FileFlow Tool

Recommendations now consider both the detected format and how it was supplied. Base64 or Data URI AVIF can open Base64 to AVIF, AVIF file paths can open AVIF to Base64, and HEIC/HEIF bytes can open Bytes to Image while HEIC/HEIF file paths can open HEIC to PDF.

Structured web data can open JSON Formatter, JWT Decoder, XML Formatter, HTML Decode, URL Decode, Base64 Analyzer, Bytes to Image or the Text Tools hub.

Private, signature-based browser detection

Detection runs locally and does not upload the pasted input. The result combines MIME declarations, syntax checks, decoded magic bytes and confidence scoring.

Format detection is a best-effort recommendation. Files with missing headers, encrypted content, proprietary containers or incomplete data may still require the Base64 Analyzer or a specialised desktop application.

AI Detect Base64, JSON, JWT and File Formats FAQ

What formats can this detector recognize?

It recognizes Base64 and Data URIs, PNG, JPEG, WebP, AVIF, HEIC/HEIF still images and sequences, PDF and Office signatures, byte arrays, hexadecimal bytes, JSON, JWT, XML, SVG, HTML, CSS, URLs, URL-encoded text, fonts, archives and several common media containers.

Can it identify a file hidden inside Base64?

Yes. For manageable inputs it decodes the Base64 locally and checks known magic bytes for image, PDF, archive, Office, font and media formats. Large inputs are identified structurally without loading the full payload into memory.

Does it detect DOCX, XLSX and PPTX?

It can identify their declared MIME types and, when enough ZIP container data is available, look for Word, Excel or PowerPoint package entries.

Is my pasted data uploaded?

No. Detection runs locally in your browser and does not upload pasted data to a server for analysis.

What happens after a format is detected?

The detector shows a confidence score, explains the matching signal, lists secondary possibilities and links to the most relevant FileFlow Tool converter or formatter.

Can format detection be wrong?

Yes. The detector is rule-based and works best when valid headers, MIME metadata or complete structured text are present. Truncated, encrypted or proprietary data may return a lower-confidence or generic result.

Technical reference

How browser-side format detection works

The detector classifies pasted data using syntax, declared MIME metadata and decoded byte signatures. It can recognize Base64 and Data URLs, byte arrays, hexadecimal bytes, JSON, JWT, XML, SVG, HTML, CSS, URLs and several common binary file containers without uploading the input.

When Base64 can be decoded safely, the tool checks magic bytes and container markers for formats such as PNG, JPEG, GIF, WebP, PDF and ZIP-based Office documents. Confidence is higher when multiple signals agree, while truncated, encrypted or proprietary payloads may remain generic or uncertain.

Signals used for detection

  • Text syntax such as JSON, JWT, XML, HTML, CSS and URLs.
  • Data URL MIME declarations and Base64 structure.
  • Decoded magic bytes and container signatures when available.

What stays in the browser

  • Pasted input is analyzed locally rather than uploaded for classification.
  • Large payloads can be identified structurally without decoding the whole value.
  • The result links to the most relevant converter, analyzer or formatter.
PNG: 89 50 4E 47JPEG: FF D8 FFPDF: %PDF-ZIP/Office: PK