Base64 to File Decoder (PDF, Image, ZIP)

Continue in the tool

More pages for this tool

invalid base64 string error

Invalid Base64 error: how to find the wrong character

Paste the Base64 that failed. The tool applies the safe fixes by itself and, for whatever is left, lists each problem with its exact position and a snippet of the surrounding text, so you can see the error at a glance.

The most common causes of Base64 that will not decode

Almost every error comes from one of these causes. First, characters outside the alphabet: quotes copied along, a comma, an accent, an invisible zero-width space or the U+00A0 the browser uses in place of a regular space. Second, the intermediate encoding: the Base64 went through a URL parameter and the + and = signs became %2B and %3D, or through a JSON string and the slashes became \/ and the line breaks became \n. Third, the length: the count of useful characters cannot leave a remainder of 1 when divided by 4, because a single character carries 6 bits, less than a byte.
The tool fixes by itself what does not change the meaning: spaces and line breaks are removed, missing padding is restored, the URL-safe alphabet is accepted, URL and JSON escapes are undone and PEM blocks lose their BEGIN and END lines. What is left is reported with the position counted from 1, the line, the column, the character and its Unicode code, plus a snippet of the text around it, where spaces appear as dots. An = sign in the middle of the text is handled separately, and it usually means two Base64 strings were glued together.
If the error still persists, there is the tolerant mode, which ignores invalid characters and drops the incomplete remainder. It exists to recover as much as possible from a damaged text, but the file may come out corrupted: one lost character in the middle shifts every following byte. Check the detected type and size afterwards. If the type is recognized, the initial part is intact; if it is unknown, the text started wrong. In code, removing the line breaks and using the right variant (standard or URL-safe) before decoding is usually enough.

Checked errors and fixes

Invalid character

Input: Zm9v\nYm@r
Expected output: Erro na posição 8 (linha 2, coluna 3): @

The position counts from 1 and includes the line break.

Impossible remainder

Input: Zm9vY
Expected output: Erro de tamanho: sobra 1 caractere

5 useful characters leave a remainder of 1; tolerant mode drops the last one.

Equals in the middle

Input: Zm8=Zm9v
Expected output: Erro: sinal = na posição 5

Two Base64 strings glued together; split them and decode one at a time.

Missing padding

Input: Zm9vYg
Expected output: foob

Fixed automatically, the = is added.

URL-encoded text

Input: %2B%2F8%3D
Expected output: bytes FB FF

%2B is +, %2F is /, %3D is =; it becomes +/8=.

Full tool FAQ

How do I convert Base64 to a PDF?

Paste the Base64 (with or without the data:application/pdf;base64, prefix) and check that the detected type is PDF. A PDF in Base64 starts with JVBERi, which is %PDF- encoded. The tool shows the PDF in the preview and downloads it with a .pdf extension. If the API response is JSON, paste the whole JSON: it finds the longest field that looks like Base64 and lets you switch fields.

How can I tell what type of file an unknown Base64 is?

What is the difference between Base64, base64url and a data URI?

What does the invalid character error mean?

Does Base64 work without the equals sign at the end?

How much does Base64 increase the file size?

Is the preview safe with unknown files?

How large can the input be?

Does it unzip GZIP or open the ZIP?

Is my file sent to any server?

Frequently asked questions

Why does my Base64 work in one place and fail in another?

Libraries differ in strictness: some ignore line breaks and missing padding, others reject them. The alphabet changes too, and + / are not accepted by strict URL-safe decoders. The tool shows which alphabet was detected.

What is the U+00A0 that shows up in the error?

It is the non-breaking space, common in text copied from web pages and e-mails. The tool treats it as whitespace and removes it without error; other invisible spaces, such as the zero-width one, are reported with their code.

Is tolerant mode safe for any file?

No. It is for recovering data from a damaged text, and the result may be corrupted. Use it on text files and images to look at, not on documents you will sign or software you will install.

Next steps

The most common causes of Base64 that will not decode

Almost every error comes from one of these causes. First, characters outside the alphabet: quotes copied along, a comma, an accent, an invisible zero-width space or the U+00A0 the browser uses in place of a regular space. Second, the intermediate encoding: the Base64 went through a URL parameter and the + and = signs became %2B and %3D, or through a JSON string and the slashes became \/ and the line breaks became \n. Third, the length: the count of useful characters cannot leave a remainder of 1 when divided by 4, because a single character carries 6 bits, less than a byte.

The tool fixes by itself what does not change the meaning: spaces and line breaks are removed, missing padding is restored, the URL-safe alphabet is accepted, URL and JSON escapes are undone and PEM blocks lose their BEGIN and END lines. What is left is reported with the position counted from 1, the line, the column, the character and its Unicode code, plus a snippet of the text around it, where spaces appear as dots. An = sign in the middle of the text is handled separately, and it usually means two Base64 strings were glued together.

If the error still persists, there is the tolerant mode, which ignores invalid characters and drops the incomplete remainder. It exists to recover as much as possible from a damaged text, but the file may come out corrupted: one lost character in the middle shifts every following byte. Check the detected type and size afterwards. If the type is recognized, the initial part is intact; if it is unknown, the text started wrong. In code, removing the line breaks and using the right variant (standard or URL-safe) before decoding is usually enough.

Checked errors and fixes

Invalid character

Input: Zm9v\nYm@r
Expected output: Erro na posição 8 (linha 2, coluna 3): @

The position counts from 1 and includes the line break.

Impossible remainder

Input: Zm9vY
Expected output: Erro de tamanho: sobra 1 caractere

5 useful characters leave a remainder of 1; tolerant mode drops the last one.

Equals in the middle

Input: Zm8=Zm9v
Expected output: Erro: sinal = na posição 5

Two Base64 strings glued together; split them and decode one at a time.

Missing padding

Input: Zm9vYg
Expected output: foob

Fixed automatically, the = is added.

URL-encoded text

Input: %2B%2F8%3D
Expected output: bytes FB FF

%2B is +, %2F is /, %3D is =; it becomes +/8=.

Full tool FAQ

How do I convert Base64 to a PDF?

How can I tell what type of file an unknown Base64 is?

What is the difference between Base64, base64url and a data URI?

What does the invalid character error mean?

Does Base64 work without the equals sign at the end?

How much does Base64 increase the file size?

Is the preview safe with unknown files?

How large can the input be?

Does it unzip GZIP or open the ZIP?

Is my file sent to any server?

Frequently asked questions

Why does my Base64 work in one place and fail in another?

What is the U+00A0 that shows up in the error?

Is tolerant mode safe for any file?

No. It is for recovering data from a damaged text, and the result may be corrupted. Use it on text files and images to look at, not on documents you will sign or software you will install.