> For the complete documentation index, see [llms.txt](https://docs.thunkable.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.thunkable.com/snap-to-place/image-recognizer.md).

# Image Recognizer

This component is powered by [Microsoft Azure](https://blogs.microsoft.com/ai/azure-image-captioning/). You can get your own server URL and subscription key for Microsoft Azure [here](https://azure.microsoft.com/en-us/free/). You can add these details to your app in the [Project Settings panel](/snap-to-place/projects/settings.md#api-keys), which you can access by clicking on your app icon.

{% embed url="<https://youtu.be/yG9mdhn70w4>" %}

## Blocks

### Upload

![](/files/-MYdUWtPSMmZq_8MFWND)

#### Inputs

| Name  | Data Type |
| ----- | --------- |
| image | Image     |

#### Outputs

| Name            | Data Type | Data                                                                                  |
| --------------- | --------- | ------------------------------------------------------------------------------------- |
| tags            | List      | Tags used to categorize the image                                                     |
| description     | Text      | Description of the image                                                              |
| confidenceLevel | Decimal   | Value between 0 and 1 describing confidence that the description and tags are correct |
| error           | Text      | If error, returns error; else returns `null`                                          |