Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform
If you want to use a custom lexicon URL in the Foundry portal, a few basic things need to be in place for it to work properly.
First, make sure the lexicon file is in the correct format and meant for speech services. Custom lexicons are used only for speech scenarios like speech to text or text to speech, and they won’t apply outside of that.
Next, check that the lexicon file is hosted at a public HTTPS URL. The service must be able to access the file without login or extra authentication. If the file is in Blob Storage, it should be public or shared using a valid access link.
Also confirm that you are adding the lexicon in the right place. Lexicons are applied as part of the speech configuration, not directly at the Foundry project or endpoint level. If it is attached to the wrong resource, it will be ignored.
Make sure the language in the lexicon matches the language you are using for speech. If the languages do not match, the lexicon will not take effect.
To validate everything, try testing with a very small lexicon and a simple speech sample. This makes it easier to confirm whether the lexicon is being applied correctly.
Most issues with custom lexicons come down to URL access, file format, or language mismatch. Once those are aligned, the lexicon should start working as expected.
Hope this helps, and feel free to share what you notice after trying these steps.
If these steps resolve your issue, please mark this answer as Accepted so others can easily find the solution. If you still face problems, share additional details and we’ll assist further.
Thankyou!