AzureOpenAI Assistant - Can't read uploaded files

[magnurgit]

I'm fairly new to Librechat, and setting this up has been a yourney, let me tell you!
My big focus in all this was to get it working with AzureOpenAI instead og regular OpenAI.
I've finally firgured that stuff out, and spun up some models via our Azure pages.
The regular models are working just fine, but i'm having a hard time gettings assistants to be able to read files.
It seems like the files are not stored properly, or that the assistant does not have access to them because of read/write properties.
I'm not sure to be honest.

Can anyone give me some tips. I'm stuck!

I'm running Librechat in Docker via a Digital Ocean server.
I would paste my docker logs, but there ain't much going on there while using the assistant and uploading files.
The best clues i've found was talking to the assistant itself:

It keeps saying this:
`# Let's take a look at the uploaded files to identify the content for review and correction.
msearch()

Let's fetch the list of uploaded files to identify the document provided by the user.

uploaded_files = !ls /mnt/data
uploaded_files`

And than tells me it cannot find the file i've uploaded.

This is my librechat.yml:

version: 1.0.5
cache: true

endpoints:
  azureAssistants:
    capabilities: ["code_interpreter", "actions", "tools", "retrieval"]
  azureOpenAI:
    assistants: true
    titleConvo: true
    titleModel: gpt-35-turbo

    groups:
      - group: "groupOne"
        apiKey: "${API_KEY}"
        instanceName: "myInstance"
        version: "2024-12-01-preview"
        assistants: true

        models:
          gpt-4o:
            deploymentName: gpt-4o
            version: "2024-12-01-preview"
          gpt-4.1:
            deploymentName: gpt-4.1
            version: "2024-12-01-preview"
          o3-mini:
            deploymentName: o3-mini
            version: "2024-12-01-preview"
          gpt-4o-mini:
            deploymentName: gpt-4o-mini
            version: "2024-12-01-preview"
          gpt-35-turbo:
            deploymentName: gpt-35-turbo
            version: "2024-12-01-preview"

If you need additional information, tell me and i'll keep an eye on this discussion to update when needed.

[magnurgit]

The only thing i've found inside the docker logs that seems to be connected to this issue is the following:
2025-06-11 08:40:05 error: [/files] Error processing file: 404 Resource not found

[magnurgit]

While doing some more research i've concluded that this is a RAG issue. I've not set that up correctly, but I cannot find any documentation on how to do this for Azure Open Ai, only for regular Open AI. Everything in the .env when it comes to Azure seems deprecated, so I presume a setup for this happens in the librechat.yml, but cannot find a single example of this.

Have anyone else solved this?

[jimmyland22]

Are you sure you're actually using the Azure Assistants you created in Azure? Last I checked it's not supported anymore.

[magnurgit]

Yes, i'm pretty sure. The only two endpoints I have activated in .env is:
ENDPOINTS=azureOpenAI, azureAssistants

And the only api-key i've got in there is for the Azure Open AI.

The assistant is working fine, except for the fact it will not read files.