The file search feature is available as part of Azure OpenAI Assistants and enables retrieval of relevant content from uploaded files. To use this capability, ensure your files are properly uploaded, a vector store is created, and you're using the correct API version—specifically, version 2024-05-01-preview or later.
As for the image vision capability, there is no mention of an "image_vision" feature in the current context, so its availability or support cannot be confirmed at this time.
If you're still encountering issues, verify that the models and API versions in use are compatible with the features you're trying to access.
Please refer this Azure OpenAI Assistants file search tool (Preview).
Thank you!