> For the complete documentation index, see [llms.txt](https://docs.ninox.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ninox.com/getting-started/builder-getting-started/intro-to-the-ninox-ui/ninox-ai.md).

# Ninox AI

Learn when to use Ninox AI and what to include for great results.

**Ninox AI** helps you turn an idea into a usable app in minutes. From a short description, it creates a strong first version with suggested tables, fields, relationships, and an initial app setup. On the app screen, it also helps you update the current app with common structure changes. Ninox also provides **AI suggestions** during table creation, so you can get helpful field proposals even when you build step by step.

That means less time spent modeling from scratch and faster progress from day one. You get something concrete to review, test, and improve right away, so you can shape the app around your process and show early results faster.

Ninox AI works especially well when you already know your use case and want to move quickly. If you know what you need to track, but do not want to build every table and field manually, the AI assistant helps you save time, reduce setup effort, and start with confidence.

You can access the Ninox AI in these ways:

* When you open Ninox for the first time, you see **Let's create your app** on the landing screen.\
  This window appears only the first time you access Ninox.
* You can also start it later from workspace **Home** by clicking **Create app with Ninox AI**.
* On the app screen, select the magic-wand **Open chat** icon in the bottom-right corner. <i class="fa-wand-magic-sparkles">:wand-magic-sparkles:</i>

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2F1bA5TBsQ3I8HAIceo923%2Fen_ai_open_chat.mp4?alt=media&token=9ce5f962-0ac8-439e-8c90-14e3b33358a3>" %}

To see the complete flow of using the Ninox AI assistant to create an app, including each step in the interface, check [Quickstart - create your first app](/getting-started/builder-getting-started/quickstart-create-your-first-app.md).

### AI assistant on the app screen <a href="#ai-assistant-on-the-app-screen" id="ai-assistant-on-the-app-screen"></a>

On the app screen, the AI assistant opens in the left panel or as a floating panel on the right. Use <i class="fa-grip-dots-vertical">:grip-dots-vertical:</i> in the upper-left corner of the chat to move it freely. Ninox remembers the last position you used.

Use the icons at the top to control the chat and editing context:

* **View chat history** shows the current chat name. Select it to **Search** chats in this app or choose one from the list.
* **New chat** starts a new chat. Use it when you want to work on a new topic separately from the current conversation.
* **Display options** lets you place the assistant in the left panel, float it on the right, or open it in full-screen mode.
* **Minimize chat** hides the chat.
* **Open chat** reopens it in the last position you used.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2FxAMIY8t9KxXq6SFlm6pJ%2Fen_ai_chat_options.mp4?alt=media&token=bae2fbb9-d4e7-48c9-8c4e-f503ffa6f43e>" %}

At the bottom of the prompt field, Ninox AI provides these options:

* The app list lets you choose which app the prompt refers to. Select one app or select all apps.\
  If you open Ninox AI while an app is open, Ninox automatically selects that app.
* Choose **Auto** to approve requests automatically. Choose **Manual** to review each suggestion, approve or reject it, or make your own suggestion. See an example recording [here](/getting-started/builder-getting-started/quickstart-create-your-first-app.md#refine-the-generated-data-model).
* The Ninox AI model selector lets you choose the model that best suits your task. Balance task complexity and speed. For example:
  * **Sonnet**: best for everyday work - fast
  * **Haiku**: best for quick questions and small edits - fastest
  * **Opus**: best for complex tasks and app building - slower
  * **GPT-6 Luna**: OpenAI for everyday changes - fast

<figure><img src="https://3739718108-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2F1fS5MB6LBvAlio8hSrnb%2Fen_ai_models.png?alt=media&amp;token=891d759a-a10b-4cab-9d73-0f7954fe0c23" alt=""><figcaption></figcaption></figure>

Below the prompt field, Ninox shows predefined prompts for common tasks. They give you a quick starting point for reviewing and analyzing your data and data structure. Use them as they are, or describe the change you want in your own words before you send the prompt.

### AI help in the logic editor <a href="#ai-help-in-the-script-editor" id="ai-help-in-the-script-editor"></a>

You can also use AI anywhere you write scripts. Open the [logic editor](/ninox-scripting/automate-your-workflows/explore-core-scripting-elements/logic-editor-features.md), describe what you want to build, and let AI help you create or refine the script.

In the logic editor, Ninox opens an AI panel next to the code area. This lets you work with AI directly while you write and refine your script.

If you are not sure how to start, click **Suggest ideas**. Ninox then shows example prompts based on the context of the current table and, for example, the logic field you are editing.\
You can select one of these suggestions or enter your own request in **Ask about Ninox script**. For the best result, mention the table, the fields, and the expected outcome. The more specific your prompt is, the closer the generated script will match your use case.\
Ninox AI then generates a script draft directly in the editor and shows a matching explanation in the AI panel.

<figure><img src="https://3739718108-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2Fc1c7hmQIeGR76Ki7Iyna%2Fscript%20with%20ai.png?alt=media&amp;token=226e9346-0e35-4a2b-832a-d3d797cff26b" alt=""><figcaption></figcaption></figure>

Review the suggested script before you keep it. You can **Accept** it, **Reject** it, or refine it by continuing the chat with Ninox AI. When you are ready to keep the current script in the logic editor, click **Apply**.

### Good habits when working with the AI assistant <a href="#good-habits-when-working-with-the-ai-assistant" id="good-habits-when-working-with-the-ai-assistant"></a>

Start with a description that gives the AI something solid to work with. Name the main things you need to track, add the most important fields, and explain how those records connect. For example, if you are planning a project app, say that each project can have many tasks and belongs to one customer.

Use screenshots and images to explain what you need. Drag and drop the files into the AI chat to attach them. Supported files are: JPEG, PNG, GIF, or WebP images.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2FIJw3Hsjojeot5IKa4jqW%2Fen_drop_file_in_ai_chat.mp4?alt=media&token=f8d80e85-e59e-4d25-b8a6-c4627cf8debc>" %}

This produces a stronger first result, but it is still only a starting point. Review the generated tables and fields, then adjust them to match your process. Before you begin, also decide which workspace the new app should belong to, so the structure fits your setup from the start.

### AI suggestions for table creation <a href="#additional-ai-feature-ai-suggestions-for-table-creation" id="additional-ai-feature-ai-suggestions-for-table-creation"></a>

In addition to the AI assistant, Ninox also offers **AI suggestions** when you create new tables. When you enter a descriptive table name and keep **AI suggestions** enabled, Ninox proposes relevant fields for your table.

<figure><img src="https://3739718108-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2FPvCF5YZzVcXbk9WGG6Yg%2FTable%20view_create%20table_AI%20suggestions_cropped.png?alt=media&amp;token=d376061f-610b-4cff-891e-77ecdfcea46e" alt=""><figcaption></figcaption></figure>

You can accept the suggested fields or turn off **AI suggestions** if you prefer to define all fields yourself.

### Query data with Ninox AI

Use natural language to find and filter records in your workspace. The **Ninox AI** lets you access your data without manually building queries. It understands your database structure, including tables, fields, relationships, and available choice options.

**Ninox AI** does not read or analyze all records in advance. It uses the database schema to find the requested information.

For example, ask: “Show me all open orders.”\
The Ninox AI identifies the relevant table and status field. It then retrieves records where the status is "Open".

**Ninox AI** uses the same filtering and search capabilities as the Ninox user interface. It interprets requests written in natural language, but cannot apply filter conditions that Ninox does not support.

{% embed url="<https://files.gitbook.com/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYwCp7NT87JGgngrkhfHT%2Fuploads%2FF83c3VMGxiGZv8zfnxKQ%2Fen_query_data_with_ai.mp4?alt=media&token=80eb2a0e-2685-465b-bf6c-1f883d077732>" %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.ninox.com/getting-started/builder-getting-started/intro-to-the-ninox-ui/ninox-ai.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.