Generate realistic fake data for testing and development. 120+ fields, 15 templates, CSV/Excel/JSON/SQL/XML/YAML export. Free, no signup, runs in your browser.
No reviews yetBe the first to leave a review for Fake Database
Maker
📌
Every developer needs realistic test data at some point. But the options are limited: Mockaroo charges $49/month for features like SQL export and Excel, Faker.js requires a backend, and random generators create isolated fields with no correlations between them.
Fake Database is a free, browser-based tool that solves this. It generates realistic fake data with 120+ fields across 19 categories, 15 pre-built templates, and 7 export formats.
What makes it different:
• Cross-field correlations — gender determines name, age determines salary, car brand determines model. Data is internally consistent, not random.
• 7 export formats — CSV, Excel (.xlsx), JSON, SQL (CREATE TABLE + INSERT), XML, YAML, and interactive table view. Most competitors charge $49/month for SQL + Excel.
• Custom columns — add unlimited fields with SQL-like types (INT, VARCHAR, DATE, UUID, etc.)
• Null & error injection — configurable rates for edge-case testing
• US + Brazilian data — locale-aware formatting for CPF/CNPJ, SSN/EIN, addresses, phones, dates
• 100% client-side — no server, no signup, no data sent anywhere
Built with React, TypeScript, and Vite. Open source.
Report
Does the data generation work fully offline once the page loads, or does it need to hit a server for any of the templates?
Report
Maker
@narinjpyw Great question, yes, it works 100% offline. Once the page loads, zero server calls are made during data generation.
Here's why:
- All generation logic runs in the browser, the entire engine (templates, dictionaries for 8 locales, correlation rules, random utilities) is bundled into the JS at build time.
- Web Worker for heavy loads, generating 1,000+ rows offloads to a Web Worker, still entirely client-side.
- Zero network requests, there are no fetch calls, no API endpoints, no backend. Every export format (CSV, JSON, SQL, XML, YAML, XLSX) is built in-memory.
- PWA with service worker, after the first visit, all assets are precached, so it even works without an internet connection.
The only external calls are Google Analytics (which silently no-ops if offline) and Google Fonts (purely cosmetic). No data ever leaves the browser.
In short: what you generate stays on your machine.
Report
Maker
@narinjpyw We just launched a Relational Mode (Beta) with multi-table generation, foreign keys, and topological sorting. All of this runs entirely client-side too, in the same Web Worker, zero server calls. Your relational data also never leaves the browser.
Report
The browser-only approach is a really thoughtful choice here, no server uploads means my test data stays where it belongs. Appreciate the export format range too, especially YAML.
Report
Maker
@ayseluzerer6pg Thanks so much, Aysel! Privacy was a non-negotiable from day one, your test data should never leave your machine, period. We're glad the export range resonates, especially YAML! It's sometimes overlooked but such a clean format for config files and CI/CD pipelines. Really appreciate you taking the time to check it out!
Report
Maker
@ayseluzerer6pg We just expanded the export capabilities with a new Relational Mode (Beta), now you can generate related tables with foreign keys and export as multi-table SQL, multi-sheet Excel, or ZIP with separate files per entity. Still 100% client-side, still no server uploads. Would love to hear what you think of the new relational exports!
Does the data generation work fully offline once the page loads, or does it need to hit a server for any of the templates?
@narinjpyw Great question, yes, it works 100% offline. Once the page loads, zero server calls are made during data generation.
Here's why:
- All generation logic runs in the browser, the entire engine (templates, dictionaries for 8 locales, correlation rules, random utilities) is bundled into the JS at build time.
- Web Worker for heavy loads, generating 1,000+ rows offloads to a Web Worker, still entirely client-side.
- Zero network requests, there are no fetch calls, no API endpoints, no backend. Every export format (CSV, JSON, SQL, XML, YAML, XLSX) is built in-memory.
- PWA with service worker, after the first visit, all assets are precached, so it even works without an internet connection.
The only external calls are Google Analytics (which silently no-ops if offline) and Google Fonts (purely cosmetic). No data ever leaves the browser.
In short: what you generate stays on your machine.
@narinjpyw We just launched a Relational Mode (Beta) with multi-table generation, foreign keys, and topological sorting. All of this runs entirely client-side too, in the same Web Worker, zero server calls. Your relational data also never leaves the browser.
The browser-only approach is a really thoughtful choice here, no server uploads means my test data stays where it belongs. Appreciate the export format range too, especially YAML.
@ayseluzerer6pg Thanks so much, Aysel! Privacy was a non-negotiable from day one, your test data should never leave your machine, period. We're glad the export range resonates, especially YAML! It's sometimes overlooked but such a clean format for config files and CI/CD pipelines. Really appreciate you taking the time to check it out!
@ayseluzerer6pg We just expanded the export capabilities with a new Relational Mode (Beta), now you can generate related tables with foreign keys and export as multi-table SQL, multi-sheet Excel, or ZIP with separate files per entity. Still 100% client-side, still no server uploads. Would love to hear what you think of the new relational exports!