API Maker
The backend engine for the AI era
API TestingEpisode 36 6:28 11 chapters November 2026
One page to call any API of your account, change its code and run it, read its logs, and take the request away as code. A tour of the API Testing page: twelve categories, query params and the model of an answer, rows generated from the schema, the answer as a table you can edit and export, keyboard shortcuts, logs, cURL and request states.
In the playlists The complete course Testing & debugging
One page lets you call any API of your account. You can change its code and run it, read its logs, and take the request away as code. In this video, we go through the API Testing page, tab by tab.
Testing a backend is usually slow. A Postman collection drifts away from the server. Tokens are copied by hand. And every try means a trip through the editor, the deploy, and the logs. With API Maker, every API of the account is listed, as the server describes it. The page signs the requests for you. And the code you test is the code on your screen. In this video: the categories and the APIs of a table, query params and the model of an answer, rows from the schema, and an answer as a table you can edit, a custom API with the keyboard, its hooks, and its logs, the request as code, and the history of your requests.
Open APIs and Logic, then API Testing. Here are its categories. The APIs of your tables, custom APIs, system APIs, and code that is not an API, like events, schedulers, migration scripts, or hooks, which you can run when you want. Pick Instance APIs. Pick postgres, then crm, then customers. You see every API of the table, schema or generated. Pick Get all.
Query Params lists what this API takes. Set limit to two, and pick three fields. Send. We get two customers, the time, and the status. Copy Interface takes the first customer of the answer. Name its model Customer, and it writes the model. Here it is in TypeScript, or in Kotlin, ready for an app. And there are many more languages.
Now pick MariaDB, the coupons, and Save single or multiple. Generate Data writes the body for you, from the schema of the table. Here is an array of two coupons, with a value of the right type in every column. Change what matters, and send it. You can make one object, an array, or as many rows as you ask for. Pick Get all, and Send. We get the ten coupons.
Switch to Table view. It is the same answer, as a grid. Use a quick filter: WELCOME5. Double click on max_uses, type twelve thousand, and Save counts one change. Click Save. The grid calls Update by id of the table, and MariaDB has the new value. Show every row again. Export gives CSV, Excel, or PDF, of the rows as they are filtered and sorted. Pick two rows, with Command click.
Delete asks first, then calls Remove by id for each row. Here, we remove nothing. You can edit a MongoDB table this way, or any table whose schema names its primary key.
Now pick Custom APIs, then Cart Quote. The Header tab signs the call with the API user Default. It is a test token, made by the server for this page only. In the body, we put a cart of a hundred and fifty dollars, with a coupon. Command E sends it. Pre Hooks shows the hook that runs before the code: Normalize cart. Logs shows what the hook and the code wrote for this call, in order.
The Code tab holds the code of the API. We add a new log line, without saving. Command E runs the code on the screen, so the new line is already in the logs. Command S saves it. Now the next call from anywhere runs it.
We send the same cart. Integrations writes this request as code, in dozens of languages and styles. Filter for the curl command. Here is the request on the screen. Copy it. It has the word token where the real token goes. Here is the same request, in Python, with requests. We paste it in a terminal, with a real token. We get the same answer: two hundred, with a total of a hundred and forty two fifty.
Every Send is kept. First the cart with the coupon, then the same cart without it, sent again. Request States shows both calls, with their headers, their body, the answer, and the code. Save the request on the screen, with a name, for you alone, or for every user, through Git.
Good to know. The Test API button on every page opens API Testing on that item. x-am-meta set to true adds what ran the call, and how long it took. The Index Maker tab lists the queries that a custom API made. Test Hooks runs one hook alone. Video thirty eight follows a whole bug with them.
The request comes from the definition of the API, not from a collection kept by hand. The code you test is the code on the screen, saved with one key. So you can check the requests an AI assistant writes here first, with the same tokens and the same logs.
To sum up. Every API and every kind of code is on one page. You get query params, and the model of an answer. You get rows from the schema, and the answer as a table, to edit, save, and export. Command E runs the code on the screen, and Command S saves it. You see the logs and hooks of the call. And you get the request as code, with its history.
Next: test cases, with mocks and coverage. API Maker. Build faster, spend smarter, deliver better.
Length 6:32Test cases
Length 7:32Test cases in depth
Length 6:46Coverage
Length 9:10Debugging