Did the webhook change — or just arrive twice?

Paste two webhook deliveries and see exactly what changed: every field added, removed or changed, every header difference, and a plain-English verdict on whether it is the same event retried or a new event.

  • Your payloads never leave the device
  • No sign-up · no install · no account
  • Same event vs new event verdict
  • Markdown + JSON reports to download

Compare two webhook deliveries

APayload A — first delivery

The delivery you received first. Paste the raw body, or drop a .json file onto the box.

0 / 500,000 characters

BPayload B — second delivery

The retry, redelivery, or second event to compare against A. Same format as A.

0 / 500,000 characters
Paste both deliveries above, then compare. Nothing is uploaded. Each body is compared up to 500,000 characters.

Paste two webhook deliveries above — or load the sample pair to see how the comparison works.

Show: 0 of 0 shown
Download the comparison. Markdown reads well in a pull request or incident note; JSON feeds scripts or CI.

Privacy you can verify, not just trust

Webhook payloads often carry customer data — order totals, email addresses, internal IDs — so a comparison tool that uploads them would be a liability. This one can't: both payloads are compared in your browser with JavaScript, on this device. Your webhook data never leaves the device — the only third-party requests are the site's own analytics, which count page views — and it uploads nothing, ever. Don't take our word for it:

New cross-origin requests during use: 0
  1. Press F12 to open developer tools, then the Network tab.
  2. Compare two payloads above. Watch: apart from the site's analytics, nothing new is requested — no uploads, and no payload data ever leaves your device.
  3. Extra credit: disconnect your internet after this page loads — pasting, comparing and downloading still work.

How to tell a retry from a new event

A webhook arrives twice and your handler runs twice — or the payload changed between the two copies and you no longer know which one is true. The fix is a field-by-field comparison plus the event ID. When you paste two deliveries here, your browser:

  1. Parses both bodies and diffs them field by field — added fields, removed fields, changed values, and type changes (a string where a number used to be), each with a plain-English explanation of why it matters to a handler. Key order and whitespace never count as changes.
  2. Diffs the headers separately, flagging signature headers — a re-signed delivery looks identical except for its signature, so signature changes are called out on their own instead of being mistaken for a changed payload. Status and response time are compared too: a 500 on the first attempt is the usual reason a provider retries.
  3. Reads the event ID and gives a verdict — the same ID on both sides means the same event; different IDs mean different events. Combined with whether the bodies match, that produces one of five plain verdicts, from “duplicate delivery” to “same event, different payload — treat with care”.

What makes this viewer different

The pages that top the results for webhook comparison are mostly hosted webhook services — their delivery-comparison view lives inside an account, behind a signup, and only compares deliveries that service itself captured. Generic JSON diff tools will diff two payloads, but they know nothing about webhooks: they cannot read an event ID, they treat a re-signed signature as just another changed string, and they say nothing about whether the two deliveries are the same event. This page is the only free, fully client-side viewer we found that compares two webhook deliveries with webhook semantics — event-ID verdicts, signature-header handling, status and timing comparison, and downloadable Markdown and JSON reports — with no account, no upload, and a live network-request counter so you can verify the privacy claim yourself.

When is comparing the right choice — and when is it the wrong one?

Comparing is right the moment a delivery arrives twice, a retry behaves differently from the first attempt, or a customer reports an event your logs show differently. It is the wrong choice as a substitute for your own logs — this tool only knows the two copies you paste, so if your handler stored a third version, check that record directly. It is also the wrong choice for verifying a signature: pasting a signature here never proves the payload came from the provider; only checking it against your shared secret, on your server, does that.

Honest limitations

It compares what you paste

If your server received something else, or the provider sent a third copy you did not paste, this diff cannot know about it. Check your own logs for the full picture.

No ID, no retry verdict

Without an event or delivery ID in the body or headers, the tool says it cannot tell a retry from a new event. It never guesses — add the ID to your logs and compare again.

Duplicate here is not proof

Matching payloads mean the two copies you pasted are the same. They do not prove your handler processed the event only once — check your idempotency keys and logs for that.

It never verifies signatures

A signature header is compared as text, not checked against your secret. Only your server, holding the shared secret, can verify a delivery really came from the provider.

Non-JSON bodies get a line diff

Form-encoded, XML or plain-text bodies are compared line by line, which is coarser than the field-by-field JSON diff. Whole lines added or removed is all it can report there.

500,000 characters per side

Very large payloads are refused rather than silently truncated. Paste the relevant excerpt, or split the body and compare the part that changed.

Frequently asked questions

Do my webhook payloads upload anywhere?

No. Both payloads are compared in your browser with JavaScript — your webhook data never leaves the device, and the only third-party requests are the site's own analytics, which count page views. You can verify it yourself with the network-request counter in the privacy section below the viewer.

How does the tool tell a retry from a new event?

It looks for an event or delivery ID — an id, event_id, delivery_id or similar field in the body, or a header such as X-GitHub-Delivery or X-Stripe-Event-Id — and compares the two IDs. The same ID on both sides means the same event; different IDs mean different events. The verdict combines that with whether the bodies match: same ID plus identical bodies is a duplicate delivery, while the same ID with a changed body means the provider re-sent the event with different content.

What if my payloads have no event ID?

The tool still diffs the bodies and headers field by field, but it says plainly that it cannot tell a retry from a new event — it never guesses. Add the delivery ID to your logs (most providers send one as a header or an id field in the body) and compare again with it included.

Why did the signature change if the payload is identical?

That is expected on a retry. Providers re-sign every delivery, usually with a fresh timestamp, so a new signature header with an identical body is normal — it does not mean the payload changed. The tool flags signature headers separately so a changed signature does not look like a changed payload.

Can I compare non-JSON payloads?

Yes. Bodies that are not valid JSON — form-encoded payloads, XML, or plain text — are compared line by line instead of field by field. The line diff is coarser than the JSON diff (it reports whole lines added or removed), but it still shows what changed.

What does “same event, different payload” mean?

Both deliveries carry the same event or delivery ID, but the payload changed between them — the provider regenerated the event instead of resending the same bytes. Do not assume the first copy is the one your handler kept: review the changed fields below, check which copy your handler stored, and ask the provider which version is correct before you act on either.

Can I download the comparison?

Yes. After comparing, download the verdict and every finding as a Markdown report (readable in a pull request or incident note) or as JSON (for scripts or CI). Both downloads are generated entirely in your browser.

What does this tool NOT do?

It does not receive webhooks, replay them, or verify signatures against a secret — it only compares the two deliveries you paste. It cannot check what your server actually received or did with a delivery, and a “duplicate” verdict does not prove your handler processed the event only once. It also does not store anything: close the page and the payloads are gone.