Dummy Data Generator

Dummy data is realistic-looking fake information used during software development and testing. You need it when real user data is unavailable, inappropriate (privacy concerns), or insufficient for testing scale. Common use cases include: populating UI prototypes with realistic content, testing database queries with varied data patterns, load testing APIs with substantial payload volumes, and developing features before production data exists.

The data is pseudorandom using JavaScript's Math.random() with curated pools for realism. Names draw from 100+ common first names and 50+ last names. Addresses combine real city names with plausible street patterns. While not cryptographically random (not suitable for passwords), the distribution and combinations appear realistic for testing purposes. For identical schemas, results differ on each generation.

Yes. The generator uses context-aware generation where fields can reference previously generated values in the same record. If you include firstName, lastName, and email fields, the email automatically derives from the generated names (e.g., john.smith@gmail.com). This works for any field order—the generator resolves dependencies automatically.

The tool supports 1 to 10,000 records per generation. This limit balances practical use cases with browser performance constraints. For larger datasets, generate multiple batches or consider server-side tools. Generation time scales roughly linearly: ~100ms for 100 records, ~2s for 10,000 records depending on schema complexity.

This occurs when fields use inconsistent key names. The generator recognizes standard variants: firstName/first_name, lastName/last_name. If you use non-standard keys like fname or firstname, the email generator cannot find the context values and falls back to random generation. Use the built-in field types or standard snake_case/camelCase naming for automatic relationship detection.

Use the Array field type, set your desired length, then click "Add Field" to define the array item structure. The child fields you add represent properties of each array element. For example, an orders array with id and total children generates: [{id: "...", total: 99.99}, {id: "...", total: 45.50}, ...].

Yes. Click Save Schema in the Options section, provide a name, and your configuration persists to browser localStorage. Saved schemas appear below the template buttons and survive page reloads. To delete a saved schema, click the trash icon next to its name. Export/import via copy-paste of the generated JSON is also possible.

CSV flattening uses dot notation for nested keys. A field user with child name becomes column user.name. Arrays JSON-stringify into single cells. If columns appear merged, your spreadsheet application may be interpreting commas incorrectly—ensure UTF-8 encoding and proper CSV parsing. For deeply nested data, JSON format preserves structure better than CSV.

Generate separately for each table with appropriate schemas, then combine the output. The tool generates INSERT statements only—no CREATE TABLE definitions. Include primary keys in your schema if you need to reference them for foreign key relationships. For complex multi-table seeding with foreign keys, generate in dependency order (parent tables first).

Date generates ISO 8601 calendar dates: 2026-07-25 (YYYY-MM-DD format). DateTime generates full ISO timestamps: 2026-07-25T10:00:00.000Z including time and timezone. Both sample from 1970 to present with uniform distribution. Use Date for birthdates, anniversaries, or calendar events. Use DateTime for created_at timestamps, event logs, or precise scheduling.

Yes. When selecting the Lorem Ipsum field type, a "Words" input appears. Specify any count from 1 to 1000. The generator repeats and shuffles a 64-word Latin pool to fulfill longer requests. Output always ends with a period and capitalizes the first letter for realistic sentence appearance.

Enum fields require at least one option. Ensure you've entered comma-separated values in the options field (e.g., active, inactive, pending). Empty options or malformed input (trailing commas, no values) triggers validation errors before generation. The tool randomly selects from provided options with uniform distribution.

The Phone type generates North American format: (XXX) XXX-XXXX with valid area codes (200-999). For international formats, use the Custom Enum type with your specific number patterns, or generate component fields (country code, area code, number) and combine with an Object type.

Direct import is not supported. However, you can manually recreate any structure using the visual builder: map JSON keys to field keys, determine appropriate types (string → lorem/enum/name depending on content pattern), and replicate nested objects with the Object type. For complex existing schemas, generate sample data in JSON format, then use the TypeScript export to get matching type definitions.

Saved schemas store in localStorage with a typical 5MB quota. Large schemas with many nested fields consume more space. If quota is exceeded, the tool shows a save error and suggests deleting older schemas. Generated data is never stored—only schemas—so large generations do not impact storage.

No. The Dummy Data Generator operates 100% client-side. All generation happens in your browser using JavaScript. No network requests are made, no telemetry is collected, and no data leaves your machine. This makes it safe for generating data with sensitive patterns or proprietary structures.

The Preview generates exactly 1 record with identical logic as full generation. Differences in appearance are due to: (1) pretty-printing toggle affecting JSON formatting, (2) root type (object vs array) changing wrapper structure, or (3) random variation between generations—click "Regenerate" to see new random samples. If structural differences appear, verify your schema hasn't changed between preview and generation.

Use the Integer type with appropriate min/max for your range. For true sequential generation (1, 2, 3...), the current version uses random integers. For strict sequences, generate with UUIDs then post-process with a script, or use Integer and sort after generation. UUIDs are preferred for test data as they reveal sorting/aggregation bugs that sequential IDs might mask.

Built-in pools are English-centric. For other languages: (1) use Custom Enum fields with your own name/address pools, (2) generate English data and transform externally, or (3) use the Lorem Ipsum type which uses Latin (language-neutral). Full internationalization of all data types is planned for future versions.