Decode Detector guess
Guess whether an unknown string looks like Base64, percent-encoding, a hash, a JWT, hex, or HTML entities.
What this tool does
Logs, database columns, and API responses are full of long strings whose format is not obvious. This tool inspects the length, the character range, and structural patterns of the input and reports which formats it most closely resembles.
Each format has its own fingerprint. A 32-character hexadecimal string suggests MD5, 40 characters suggests SHA-1, and 64 characters suggests SHA-256. Three dot-separated segments made of Base64 characters indicate a JWT. Repeated percent signs followed by two hex digits indicate percent-encoding, while fragments starting with an ampersand and ending with a semicolon indicate HTML entities. A string made only of A-Z, a-z, 0-9, plus, slash, and equals whose length is a multiple of four is a Base64 candidate.
When to use it
Use it as a first step when an unknown token appears in an error log, when you need to know how a stored column was encoded, when a sample data file arrives from another team, or when deciding which decoding tool to reach for next.
Input and output examples
5d41402abc4b2a76b9719d911017c592
Likely MD5 (32 hex characters)
Detection relies on length and character range, so any other 32-character hex value looks the same.
eyJhbGciOiJIUzI1NiJ9.eyJzdWIiOiIxIn0.abc
Likely JWT (three dot-separated Base64 segments)
The structure is distinctive enough to move straight to the JWT tool.
Notes and limitations
This is a pattern-based estimator, not a definitive identification. One string can satisfy the conditions of several formats at once; a purely alphanumeric value can look like both Base64 and hexadecimal. Hashes are one-way functions, so identifying the format never lets you recover the original value. Treat the result as a hint that guides the next step.
Frequently asked questions
If it is a hash, can I recover the original?
No. Hash functions are one-way, so only the format can be identified.
Several formats are reported at once.
That is expected when the character set satisfies more than one rule. Pick the candidate that fits the context and try decoding it.