Developer tools · 0.1.0
Keep the relationships.
Change the selected fields.
Prepare related API examples for documentation and tests. Replace values, remove keys and give matching IDs the same stand-in across JSON files.
Browser localNo accountWorks offline1. Choose your input
Paste JSON or select 1–20 files, each up to 2 MB. Inputs stay in this page's memory; there is no storage, upload or automatic detection.
2. Define selected fields
Use explicit slash paths. * matches one level. Token rules sharing a namespace share IDs across this batch.
Saved rules include your replacement values and paths. Use examples rather than secrets. Parent/child conflicts and zero matches block output.
3. Inspect and export
Matched fields and actions (first 200)
The manifest contains field paths, counts and original types, but no source values or token map. Paths can themselves identify people. Tokens reveal equality and structure; this is not encryption or a guarantee of anonymity.
Rule reference
{"path":"/users/*/email","action":"replace","value":"example@example.test"}
{"path":"/orders/*/userId","action":"token","namespace":"USER"}
{"path":"/metadata/internalNote","action":"remove"}
Object keys are removed. Removing an array item writes null to preserve positions. Token nulls stay null; numeric IDs become strings. Tokens distinguish numeric 7 from string "7".
Escape ~ as ~0, / as ~1 and a literal * key as ~2. The asterisk extension is specific to this tool. A root-wide rule is not supported.
Scope and limits
No automatic sensitive-data detection, synthetic generation, live API, database connection or regulated-data certification. Use approved inputs and review all unchanged values.
JSON objects/arrays only; duplicate keys, unsafe integers, invalid JSON and overlapping rules are rejected. Large IDs must be quoted. Fractions use JavaScript number precision.
Maximum: 20 documents; 2 MB and 50,000 nodes per document; depth 64; 100 rules; 20,000 selected matches per batch. Rules with no matches are listed; a completely unmatched batch is blocked.
Ordering controls sequential token assignment. Use one batch and the same file/rule order for reproducible tokens. Tokens reset on every run. No reversible map is saved.
Related-fixture download available
Related-fixture recipe kit
The free workbench is fully useful on its own. The experimental kit adds 12 original synthetic exercises, rules, expected outputs, review notes and a dependency-free Node batch runner. The test price is $9; paid demand is not yet validated.
No purchase or payment is available yet. See the free walkthrough and try the paired users/orders example first.
Common questions
Does my JSON leave this page?
The tool code has no fetch, telemetry or storage calls. File reads and transformations happen locally. Loading the hosted page is an ordinary website request; use the downloaded single HTML file for offline work.
Why use this instead of a free library?
Use Pino's redaction tools for application logs, Presidio for detection workflows, or Mockaroo for new synthetic data. This workbench offers a no-install visual review of selected fields in several related JSON files. It does not replace those broader tools.
Can I use the output in production?
Outputs are prepared examples. You decide whether they satisfy your use case. Review field names, unselected text and accidental identifiers, and keep originals in approved private storage.
Follow a worked free exercise
Mask related JSON fixtures without losing example ID relationships. Run synthetic proof and download the sample.