Base Decoder
Auto-detect Base16/32/45/58/64/85 encodings and peel nested layers until readable text or a known binary signature appears. Runs locally, no upload.
Decode unknown base-encoded text
Paste an encoded blob and this tool tries the common base encodings — Base16 (hex), Base32, Base45, Base58, Base64 (standard and URL-safe), and Base85 — then keeps peeling if the decoded result is another readable encoded layer. The report shows the detected chain and the final text, or a binary signature plus a hex preview when the bytes look like a file. Everything runs locally in your browser.
Worked example
Input:
U0dWc2JHOHNJRmR2Y214a0lRPT0=
Output:
Detected 2 layer(s): base64 → base64
Hello, World!
The first Base64 decode produces another Base64 string, so the decoder peels one more layer. Use Output style = plain when you want only the final decoded value without the layer report.
Options
- Maximum layers to peel — stops recursive decoding after this many accepted layers. The default is 8 and values are clamped to 1–30.
- Output style —
reportexplains what was detected;plainreturns only the final text, or hex when the final bytes are binary.
Limits & edge cases
- Inputs are capped at 1 MB of encoded text.
- Auto-detection is heuristic: ambiguous strings are tried in a fixed order and only accepted when the decoded bytes are mostly printable text or match a known binary signature.
- For a single known scheme where every option matters, use the dedicated codec tools (Base32, Base58, Base64, Base85, hex) instead.
- Encrypted or compressed data is not decrypted/decompressed; this only removes base-encoding layers.
Which encodings are detected?
Base16/hex, Base32, Base45, Base58, Base64 (standard and URL-safe, padded or unpadded), and Ascii85/Base85. Whitespace is ignored for most bases, so copied wrapped lines usually work.
Can it decode several nested layers?
Yes. After each successful decode, the result is tested again as text. If it looks like another supported base encoding, the next layer is peeled until the depth cap, plaintext, or a recognized binary signature is reached.
What happens when the decoded bytes are binary?
If the output starts with a known magic byte signature (for example PNG, JPEG, PDF, ZIP, gzip, ELF, or MP3), the report names that signature and shows a hex preview. Plain mode returns the final bytes as hex.
Why did it leave my input unchanged?
The text may already be plaintext, may use an unsupported alphabet, may be too short to distinguish safely, or may decode into non-printable bytes without a known file signature. The tool avoids guessing when the candidate does not look like a real decoded layer.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool base-decoder "U0dWc2JHOHNJRmR2Y214a0lRPT0="New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/base-decoder/?input=U0dWc2JHOHNJRmR2Y214a0lRPT0%3D&max_depth=8&output=reportMachine-readable descriptor: tool.json — title + parameters JSON Schema for agents.