# Custom Microsoft Image Recognizer?

**URL:** <https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020>\
**Category:** Questions about Thunkable X\
**Created:** [June 17, 2018, 10:23pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020 "2018-06-17T22:23:23Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![James\_Colestock](https://sea1.discourse-cdn.com/flex015/user_avatar/community.thunkable.com/james_colestock/32/20682_2.png) [@James\_Colestock](https://community.thunkable.com/u/James_Colestock)\
**Post date:** [June 17, 2018, 10:23pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020/1 "2018-06-17T22:23:24Z")

</div>

Has anyone successfully **overwritten the default settings for the Microsoft Image Recognizer** in #thunkableclassic or #thunkablex. When using this feature in the classroom, I have had struggled to override it. To insure that all student apps work, I want to provide my own URL and key. Has anyone done this?

![image-recognizer-properties](https://us1.discourse-cdn.com/flex015/uploads/thunkable/original/3X/5/5/5567a7bfb47c7bd87e5e23cfff85837edd3c4b4b.PNG)

What should the `ServerURL` be changed to; is it the endpoint, e.g., `https://westus.api.cognitive.microsoft.com/vision/v1.0`

What about the `SubscriptionKey`? Is it the one listed in Azure, like so:

![azure-keys-blurred](https://us1.discourse-cdn.com/flex015/uploads/thunkable/original/3X/4/0/40bfeb6fe8ab74464b0f8f32e7d13c7343460db7.png)

Thanks,

James

---

<div class="post-metadata">

**Author:** ![leo.armesto](https://avatars.discourse-cdn.com/v4/letter/l/f9ae1b/32.png) [@leo.armesto](https://community.thunkable.com/u/leo.armesto)\
**Post date:** [March 2, 2021, 12:29pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020/2 "2021-03-02T12:29:50Z")

</div>

I also have the same problem. It seems that the image recognizer is not working properly, so I tried to set my own server. Actually, I was able to check that the server side is correct, by uploading my own images and getting a description using the following cURL command:

curl -H "Ocp-Apim-Subscription-Key: " -H “Content-Type: application/json” “[https://westeurope.api.cognitive.microsoft.com/vision/v3.1/analyze?visualFeatures=Categories,Description&details=Landmarks](https://westeurope.api.cognitive.microsoft.com/vision/v3.1/analyze?visualFeatures=Categories,Description&details=Landmarks)” -d "{“url”:"[http://upload.wikimedia.org/wikipedia/commons/3/3c/Shaki\_waterfall.jpg"}"](http://upload.wikimedia.org/wikipedia/commons/3/3c/Shaki_waterfall.jpg%22%7D%22)

I tried to set the serverURL to “[https://westeurope.api.cognitive.microsoft.com/vision/v3.1/analyze](https://westeurope.api.cognitive.microsoft.com/vision/v3.1/analyze)”, “[https://westeurope.api.cognitive.microsoft.com/vision/v3.1](https://westeurope.api.cognitive.microsoft.com/vision/v3.1)”, “[https://westeurope.api.cognitive.microsoft.com](https://westeurope.api.cognitive.microsoft.com)”, and many more combinations, but none of them worked for me. Please, help me!

---

<div class="post-metadata">

**Author:** ![trawellit](https://sea1.discourse-cdn.com/flex015/user_avatar/community.thunkable.com/trawellit/32/66464_2.png) [@trawellit](https://community.thunkable.com/u/trawellit)\
**Post date:** [October 3, 2021, 7:48pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020/3 "2021-10-03T19:48:29Z")

</div>

Hi Leo,  
I would like to have my own library of image too (specific churches, monument that the app could recognize).

Can you explane how do you did?  
Or can you give me some references about?  
Thanks

---

<div class="post-metadata">

**Author:** ![leo.armesto](https://avatars.discourse-cdn.com/v4/letter/l/f9ae1b/32.png) [@leo.armesto](https://community.thunkable.com/u/leo.armesto)\
**Post date:** [October 3, 2021, 9:19pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020/4 "2021-10-03T21:19:18Z")

</div>

Apparently something was wrong with the code I used my first time (maybe the servers were down). I was able use this feature in a conventional way by redoing all the app from scratch, I would say exactly as it was before. Unfortunately this is not very informative, I know,… but I can’t tell why it was not working at the beginning.

---

<div class="post-metadata">

**Author:** ![ioannis](https://sea1.discourse-cdn.com/flex015/user_avatar/community.thunkable.com/ioannis/32/146956_2.png) [@ioannis](https://community.thunkable.com/u/ioannis)\
**Post date:** [November 8, 2024, 12:02pm UTC](https://community.thunkable.com/t/custom-microsoft-image-recognizer/32020/5 "2024-11-08T12:02:24Z")

</div>


