How to check texts for AI directly in your own system

When you check texts for AI regularly, some of the work is simply moving them between services. You open the detector, paste a text, wait for the result, then save it with the right document. With an AI detector API, the system you already work in can do this.

It's AI receives text from your application and returns the result of the check. After integration, you can show the score next to the document and send texts for analysis when you need to. Staff will not have to move each text to the detector by hand and bring the result back.

Single and batch checks with the It's AI API

It's AI API v2 has two methods. The choice depends on how you receive the content. You can check each text separately or send a group in one request.

When texts arrive one at a time

A single-text check is suitable if you want a result as each document arrives. For example, someone submits text through a form, and your application sends it to It's AI. A developer can also add a “Check” button directly to the document page.

This uses the POST /api/v2/text method. After analysis, the result returns to your application, where you can display it or save it with the document.

Each request is processed separately. You can use a score as soon as you receive it, without waiting for results on other texts.

When it is easier to check a group of texts

If the texts are already collected, you can send them through POST /api/v2/batch. This is suitable, for example, for documents uploaded in one batch or for checking accumulated content on a schedule.

On Enterprise, one request can contain up to 250 texts. You can send a whole collection this way, without making a separate request for each text.

Each text will have its own result. The results come back in the same order as the submitted texts, so the application can save each score with the corresponding document.

Send the texts as separate items in the batch. You will then get a result for each one, rather than a single score for the combined text.

Both methods are described in the It's AI API v2 documentation. You can also check several documents through separate requests: choose the approach that suits your process.

What you can show next to a document

The API returns data that a developer can use to display the result the way you need it. You can keep the overall score on the document page and show the details when someone opens the report.

Source: API response documentation.

For example, a list of documents can show the AI percentage, while the page for a selected text shows sentence scores. A person can look at the score for each sentence and find passages with high AI scores. They do not need to move the text to another service to do this: a developer can build the detailed view directly in your application.

When setting this up, keep the meaning of the measures clear. Use ai_percentage for the percentage, while score gives the detector's overall assessment. The developer sets up highlighting in your interface using the segments data.

How to share the result with someone else

The result may be needed by someone other than the person who ran the analysis. For example, a client may ask for a report along with the finished work. Through the API, you can get a PDF certificate of the check and attach it to the document.

This uses the report_id returned with the result. The certificate has a QR code linking to the report page. According to the official SDK description, the recipient can open that page through the link. You can share the report itself instead of copying the score into a message.

You decide when to run the check

The API lets you connect analysis to an action in your application. This could be a button click or the arrival of a new text. If you prefer to collect the texts first, a developer can set up scheduled requests.

For example, an employee adds a document to the system. It sends the text to It's AI, receives the response, and writes the score to the document record. When the employee opens the completed result, they no longer need to find a separate report and move the data by hand.

The request can run in a background job. While the analysis is running, the user continues working in the application; after receiving the response, the system updates the status and shows the result.

It is worth saving the score with the version you sent for analysis. As work on the document continues, you will be able to see which text the check refers to.

What other checks you can add

It's AI also offers plagiarism and grammar checks through the API. The official SDK has separate methods for them. The plagiarism check returns sources and matching passages, while the grammar check returns comments and suggested corrections. If your team needs these results too, a developer can add them to the same system by setting up separate requests.

How to start using It's AI API

You will need a developer and an API key to set up the integration. According to the documentation, API access is included in the Enterprise plan. The minimum length of one text is 200 characters; the maximum depends on the plan's terms.

Before connecting, you just need to decide:

• Where the application will get the texts.
• When they should be sent for checking.
• Where the employee or user should see the result.

These decisions will help the developer choose the method and set up the data transfer. For Python applications, there is an official It's AI SDK with synchronous and asynchronous clients.

Start with a few of your own documents and look at the whole process in your usual interface: sending the text for analysis, receiving the score, and saving it. This will help you decide which actions your team can hand over to the system in its regular work.

To prepare the integration, give your developer the It's AI API documentation, an example document, and an indication of where you want to see the result of the check.

FAQ