Building is similar to building an environmental system
Each software project is similar to building a small, live ecosystem. You just write a symbol – you can identify relationships, behaviors, structures, flows and restrictions. Everything affects everything else: The database model informs the application programming interface, which nourishes the front end, which affects UX decisions, which are due to health verification rules and error.
In the middle of that network, there is a quiet but decisive demand: Data. Each part of the stack needs – not only in the end, but from the beginning.
The moment when initial models, testing or flowing design begin, you need something to navigate the system. You need fake data. Not only for one field or one example-but complete groups of realistic organisms, organized, compatible with rules that appear and behave like inputs in the real world.
It appears at every stage:
- The user interface design with real appearance content
- Logic test that depends on field restrictions
- Simulation of API responses during integration
- By planting local environments before there is a back logic
- Show flows with actual diversity, not brand values
But generating this data is usually annoying. Most of us are either manually fake, repeat a handful of examples, or tend to tools that do not understand the structure of the chart or the level of the field level. The process works – until it is broken. Then it is broken often.
Fake data is easy – so that you are not so
Developers, quality assurance engineers and heavy data team teams hit the same wall at some point: the need for fake, realistic-fast-rapid organized data. It is not difficult to ridicule a single user object. What feels frustrated is the mockery of the multiple records of users who have valid email messages, local names, organized descriptive data, restrictions such as length, range, style or interrelated field relationships.
Mockingjar
“It is a tool built of this pain.”
Not the startup field. It is not another stadium for random fake data. Just a solid web application designed to create test data that suits your scheme, follows the rules, respects the structure.
It combines the creator of the optical chart, a strong health verification engine, and the background of a generation of AI using API Clade. You define the shape. You can direct the content. Mockingjar fills the rest – intelligently, intelligently and without requesting manual repairs.
Why this exists
This project came from repetition. Over and over, start with:
“Let’s build a quick fake for this.” After a few hours, JSON is still writing by hand, patching fake values, or adjusting knowledge and sects that do not fully match the real structure.
There is mockingjar to get rid of this friction – especially in the early stages of construction programs. These days, engineers spend more time in initial models, check ideas, and build POCS more than ever. With this pace, the need for fake, clear, organized data in the chart is not a state of edge-it is a daily life.
This tool will return you again at the time – by doing one thing well: create organized data from the scheme that you specify.
What do you actually do?
Mockingjar is divided into two main components:
1. Web interface
The full feature Web app is designed with:
- Next.js 15 (Application router)
- V7 user interface materials
- Reaction Figure + zod To verify
- The following authentication + Google Oau To log in safe entry
- CSRF + Session + Protected Roads For multiple layers of security
- Postgresql Via PRISMA for continuing data
- Zustand + React Query To synchronize the state of the front end and data
This interface includes:
- The creator of the optical scheme with a tree width
- The interface is a natural language to generate the data that AI drives
- Personal plan library with support/editing/deleting support
- Approving and managing the session
- Json preview tools and export
2. Mockingjar-Lib Core Engine
Independent Typescript package with:
- The creator of the plan and the manipulation engine
- Json auditor at the field level
- Claude Anthropor Integration Layer
- Supporting overlapping objects, rows, references and complex restrictions
- Surgical renewal of unique fields only
- Full test coverage
This is the same engine used in the web app – also usable in background textual software, or the test contestants.
Planning building
Mockingjar revolves around the user knowledge. You can specify the types of logical and brown fields and restrictions to each level.
Includes supported species:
textandnumberandbooleananddateandemailandurlobject(With overlapping fields)array(With the restrictions of elements and intertwined structures)schema(A reference to another preserved scheme)
Each field supports logic, such as:
- Wanted/optional
- Minimum values or lengths/maximum
- Match pattern (for example
^ORD-\d{6}$)) - Census options
- Size restrictions for sects
Planning composition
You can include the plans within others. The “plan” type allows you to define shared models – such as an inclusion Address Scheme in multiple forms (shippingAddressand billingAddressEtc.) without redefining it every time.
The construction scheme stands out
- String
- Tree view of deep overlap
- JSON-To-Schema (Paste in Json’s Paste and convert it into a chart)
- Square selection dialog (to include current charts)
- Shared molds
Data generation
Once the scheme is ready, you can create data with a claim like:
“Create the request data for a customer based in London, and buy a product linked to home facilities (for example, kitchen devices, cleaning equipment or home maintenance tools)“
The wave has not replaced your plan – it simply directs the tone and realistic the resulting values. Mockingjar Claude Sonit 4 is behind the scenes to fill each field during respect:
- He writes
- Restrictions
- Nesting
- Cover lengths
- Regex and Enum rules
You can configure the number of records you want.
You will get a preview result and take out coordination. If any fields fail to verify health, the health will be reserved individually – they have not been brought up with the rest. This recovery process is the key to maintaining data generation quickly and consistently.
the mockingjar-lib The primary engine
This is where all real magic lives – and you can use it directly in Node.js projects with any preferences for the interface.
example:
import { Generator } from 'mockingjar-lib';
const result = await Generator.generate(
'your-anthropic-api-key',
yourSchema,
'Generate Indonesian customer records with valid emails',
{
count: 10,
enableFallback: true,
timeout: 60000
}
);
if (result.success) {
console.log(result.data);
} else {
console.error(result.errors);
}
You can also:
- Verify JSON’s health against the chart (
Validation.validate)) - Convert json into a chart (
Schema.convert.jsonToSchema)) - Add/update/remove the overlapping fields deeply
- Get the descriptive data organized on obstetric performance
Architecture that focuses on the developer
- Filled Typescript Codebase
- Esnt + strict mode
- A normative code base with
generatorandschemaandvalidationAndanthropicLayers - It can be easily expanded to add custom generators or field types
It is still developing – open to cooperation
Mockingjar is not a polished product. It is a real functional tool that solves a specific problem-it does it well-but there is a large space for growth.
Some areas still need to improve:
- The ability to use the user interface and form the work environment
- Plan and import/export
- More intuitive treatment of deep overlapping structures
- Improving interpretation and quick comments from artificial intelligence
- Community scheme/mold sharing
- The best vision of error and error correction flows
The basic generation engine (mockingjar-lib(Solid and testing, but the surrounding experience – especially in the application of the web – is still in formation.
This project does not try to be everything from the first day. It is built from the party you put in mind, and if you are the type who sees opportunities in this type of basis, your help will be welcome.
Mockingjar is completely open source. Contributions, testing, comments, error reports, and all PRS – whether they are working to fix the defect, improve the rapid flow, or just indicate the place of feeling something.
If this tool has even a slight inconvenience to you, and you want to help push it further, then transfer.
Let’s build something useful – together.
Final thought
Mockingjar was not made to persuade. It was made for Stop wasting time – Especially on a mission that no one wants to do more than once.
If you write a plans or settings of fake application programming facades or create test settings, this may be a tool that is worth adding to your assertion. If not, then steal the parts you love – that’s what is open.
