Single button for file-uploads

[ethux]

It would be nice to have a single button to upload file types, being able to remove the upload 'image' button and just handle the images the same way as all other documents, maybe I'm missing something but I can't find how to do this in the documentation.

If someone could point me in the right direction, that would be nice :)!

I'm using ModelSpecs to create a model

• https://www.librechat.ai/docs/configuration/librechat_yaml/object_structure/model_specs

[Pfaufisch]

Don't know about Sharepoint and it will depend on the used provider/model, but in v0.8.1, disabling File Search and not enabling context in the agent capabilities should leave you with a single "Upload to Provider" button that handles all files.

[ethux]

@Pfaufisch Hi thank for your response, this doesn't seem to work for models created with "modelSpecs".
Disabling FileSearch and context reduces the option for uploading files to image only.

My config file:

version: 1.2.8
cache: false

interface:
  customWelcome:  <redacted>
  fileSearch: false
  endpointsMenu: false
  modelSelect: false
  parameters: false
  sidePanel: false
  presets: false
  prompts: false
  bookmarks: true
  multiConvo: true
  agents: true
  runCode: false
  webSearch: false
  fileCitations: true
  mcpServers:
    placeholder: "Connecties"

registration:
  socialLogins: ["openid"]

endpoints:
  agents:
    capabilities: ["actions", "tools"]
  custom:
    - name:  <redacted>
      apiKey: "${OPENAI_API_KEY}"
      baseURL: "https://api.openai.com/v1"
      models:
        default: ["gpt-4.1"]
      fetch: false
      titleConvo: true
      titleModel:  <redacted>
      modelDisplayLabel:  <redacted>

  azureOpenAI:
    titleConvo: true
    titleModel:  <redacted>
    groups:
      - group: "Azure OpenAI Switzerland"
        apiKey: "${AZURE_OPENAI_API_KEY}"
        instanceName:  <redacted>
        version: "2024-12-01-preview"
        baseURL:  <redacted>
        models:
           <redacted>:
            deploymentName: "<redacted>"

mcpServers:
  Zoeken_op_het_internet:
    command: npx
    args: ["-y", "@oevortex/ddg_search"]

fileConfig:
  serverFileSizeLimit: 1000  
  endpoints:
    default:
      totalSizeLimit: 200     
      disabled: false
     <redacted>:                  
      fileLimit: 5
      disabled: false
      fileSizeLimit: 50
      totalSizeLimit: 200
      supportedMimeTypes:
        - application/pdf
        - text/.*
        - application/vnd.openxmlformats-officedocument.*   
        - application/msword
        - application/vnd.ms-excel
        - application/vnd.ms-powerpoint

modelSpecs:
  enforce: true
  prioritize: true
  list:
    - name: <redacted>
      label: <redacted>
      default: true
      description: <redacted>
      preset:
        capabilities: ["actions", "retrieval", "tools", "file_search"]
        endpoint: "azureOpenAI"
        model: "<redacted>"
        maxContextTokens: 128000
        max_tokens: 4096
        tools: [
          "sys__server__sys_mcp_Zoeken_op_het_internet",
          "web-search_mcp_Zoeken_op_het_internet",
          "fetch-url_mcp_Zoeken_op_het_internet",
          "url-metadata_mcp_Zoeken_op_het_internet",
          "felo-search_mcp_Zoeken_op_het_internet"
        ]
        promptPrefix: <redacted>

[Pfaufisch]

Are you running the latest version (librechat:v0.8.1-rc2)? "Upload Image" got changed to "Upload to Provider" in that version, and it should be working with Azure and your model spec.

[ethux]

@Pfaufisch Thank you! I was running v0.8.0, running the newer version worked.
Besides that is it me, or are the docs sometimes a bit confusing for specific implementations?

Right now the upload to provider button is only allowing PDFs and images, but I'm sure I have the right supportedMimeTypes configured.

[Pfaufisch]

Nah, the docs unfortunately are super confusing :/ I'm not even sure thosesupportedMimeTypes refer to the reworked Upload to Provider feature. I can upload different file types depending on the provider, so my impression is this is configured somewhere outside the librechat.yaml. Maybe @dustinhealy can help?

[ethux]

That would be nice if he could help! Didn't find a solution to upload docx files yet :/

[dustinhealy]

@ethux For the most part, you can think of Upload to Provider as Upload Image + PDF for all endpoints except Google (which also accepts audio and video inputs). We do have plans to consolidate everything into a single upload flow eventually, and Upload to Provider was one of the first steps in achieving that.

For your immediate problem, I would recommend using Upload as Text as a way to parse and inject your .docx file content into the context window if you can't / don't want to just copy paste it into the chat as plain text.