Compare API payloads, config versions and migration outputs with confidence
Shipping a feature flag rollout, reviewing a partner webhook sample or proving that a database migration did not drop fields all benefit from a structured JSON diff. Text diff tools highlight commas and whitespace noise; a semantic flattening view shows which keys appeared, disappeared or changed value regardless of pretty-print style. This page is built for data and platform engineers who need a fast answer before opening a ticket: did staging and production diverge on nested attributes, or is the parser wrong? Everything executes locally so paste redacted production samples without sending them through a third-party paste service.
Diff workflow
- Validate both documents in the JSON Validator if either side came from logs.
- Paste baseline JSON on the left and candidate JSON on the right.
- Scan added (green), removed (red) and changed (yellow) rows; hide unchanged if needed.
- Export or copy findings into your change request with the path column as the source of truth.
Flattening rules and limits
Objects recurse by key name; arrays contribute numeric index segments in the path. Values that are objects or arrays on one side but scalars on the other register as a change at the parent path. The comparator treats null, empty string and missing keys distinctly so nullable API fields do not silently match. Order of keys in the original JSON does not affect the outcome because comparison uses the flattened map, not raw text lines.
Where JSON diffs show up in real workflows
Analytics engineers diff warehouse export samples after an upstream schema bump. Backend developers compare OpenAPI example fixtures across versions. SREs validate Helm values files before a cluster upgrade. Pair this tool with the JSON Formatter when you need readable input, the JSON Validator when a paste fails to parse and the JSON Schema Generator when you want to codify the shape you expect after the diff is clean.
Review checklist before you share results
Redact secrets and personal data before pasting production payloads, even though processing stays local. Confirm both sides use the same numeric precision — a float rendered as a string can look like a breaking change. When arrays reorder without semantic impact, expect path-level churn; sort keys upstream if your team only cares about set membership. Export a short list of changed paths for reviewers instead of entire documents so tickets stay readable.
Wire diffs into CI as contract tests
The manual diff is the interactive version of something worth automating. Capture a known-good response as a golden fixture in git, then have CI compare each build's output against it and fail the pipeline when paths appear, disappear or change type unexpectedly. That turns "did we accidentally break the API contract" from a production incident into a red check on the pull request. Use this page to investigate and explain a failing diff by hand, decide whether the change was intended, then update the golden file deliberately rather than letting snapshots drift silently with every merge.