A JSON Web Token is commonly represented as three dot-separated segments: a Base64url-encoded header, a Base64url-encoded payload, and a signature segment. A decoder reverses the first two encodings so developers can inspect claim names, token type, and timing fields. This tool parses the token structure and formats decoded JSON for display. It does not validate the signature, prove who issued the token, or establish that a claim is trustworthy. The distinction is essential: Base64url is an encoding, not encryption, so readable claims are not protected from someone holding the token.
Paste a token into Paste JWT token, then select Decode JWT. The action is disabled for an empty input. The tool displays the parsed header and payload in separate formatted blocks, shows the signature segment, and converts truthy `iat` and `exp` values into human-readable ISO timestamps. If an expiration value is present, the display marks it expired or valid relative to the current time. Use Clear to reset the token, output, and error state; editing the input also clears the prior result so stale claims are not mistaken for new ones.
For example, a token shaped like `header.payload.signature` is split at its two dots; the header and payload segments are Base64url-decoded and parsed as JSON. If the payload contains an `exp` claim expressed as Unix seconds, the tool converts it to a date and compares that date with the browser's current time. Changing the signature segment does not alter what the first two segments decode to, and the displayed claims can still be forged. Therefore a decoded `role: admin` claim is not evidence that the token is valid.
The output is explanatory text and JSON, not a verified-token status. The header may identify an algorithm or type, the payload may contain registered or application-specific claims, and the signature is displayed as its original third segment. Tokens with anything other than three dot-separated parts, malformed Base64url, invalid UTF-8, or invalid JSON trigger an error. Missing or zero-valued timing fields may not be shown because the helper checks truthiness. The page does not expose cryptographic verification keys, issuer selection, audience validation, or a claim policy.
Developers use a decoder to inspect a development token, troubleshoot claim naming, check a sample expiration timestamp, and understand the compact token format. JWT Encoder creates a matching unsigned preview from JSON fields; HMAC Generator helps demonstrate a keyed MAC separately. For production verification, use a trusted backend library and validate signature, issuer, audience, expiry, allowed algorithm, and application-specific authorization rules. A decoded payload is useful debugging context, not an access-control decision.
The visible client code decodes the pasted string in the browser and does not call a JWT decoding service. Even so, a JWT can contain personal or confidential claims, and the token may itself be a bearer credential. Do not paste a live production token into a shared screen or public support transcript. Avoid sharing the signature or full token in screenshots; clear the field after inspection. Local decoding does not prevent browser extensions, screen capture, or other local observers from seeing displayed text.
Treat the tool's expiration label as a display aid, not a complete validity verdict. It depends on the token's `exp` claim being in seconds and the local clock being reasonably correct; it does not check revocation, key rotation, issuer policy, or signature integrity. If parsing fails, check for copied whitespace, missing segments, or damaged URL-safe Base64 characters. Use synthetic values for demonstrations and verify any real token only in an authorized environment with server-side cryptographic checks.
JWT claims are simply JSON data carried in the token. Registered names such as `sub`, `iss`, `aud`, `exp`, and `iat` have conventional meanings, but the receiving application decides which claims to require and how to interpret them. A token can omit claims, contain custom fields, or use unexpected types. The decoder displays parsed JSON without applying a schema, so compare the result with your service's documented contract instead of assuming every field is present or trustworthy.
A useful troubleshooting approach is to inspect the header for the declared token type and algorithm, then review only the claim fields needed for the issue at hand. Avoid copying the complete token into bug trackers because it may remain usable as a bearer credential until it expires or is revoked. For a reproducible report, replace user identifiers and secrets with synthetic values while preserving the claim structure. If a timestamp looks wrong, confirm whether the producer encoded Unix seconds rather than milliseconds.
A token may be opaque to a browser application even when its structure is readable. Many clients should pass it to the authorized service without attempting to interpret claims, while the service validates it under a defined trust policy. Do not change application behavior based solely on a client-side decoded role or permission value. If a backend rejects the token, collect the server's safe diagnostic information and verify signature, issuer, audience, clock, and key configuration in a controlled environment.
When copying an example into a test, preserve the three segments and avoid trimming the trailing signature or adding punctuation. Redact subject identifiers before sharing screenshots, and use a fake issuer or audience when the page is for a demo. If you need a trustworthy token verdict, consult the issuing service's server logs or a controlled verifier that has the correct public key and validation settings.