OnlyOffice AI support

[Hello-World-Traveler]

Is your feature request related to a problem? Please describe. Using Ollama and openAI option WRN Client error ip=1 latency="121.604µs" method=OPTIONS status=401 url=/v1/models

https://www.onlyoffice.com/ai-assistants.aspx

Incorrect HTTP Method: The endpoint /v1/model may not support the method (e.g., OPTIONS) being used—many AI API endpoints only allow GET or POST.

Misconfigured AI Provider Integration: The method sent by ONLYOFFICE or another client does not match the LocalAI API specification for that route curl -X OPTIONS http://IPv1/model -H "Authorization: Bearer key" 405 Not Allowed

Describe the solution you'd like

JS file

class Provider extends AI.Provider {
  constructor() {
    super(
      "LocalAI",
      "http://localhost:8080/v1", // change to your LocalAI endpoint
      "yourLocalAIAPIKey",        // your key if applicable
      "v1"                        // model/API version as required
    );
  }
}

Onlyoffice to connect to LocalAI

[Splarkszter]

Same issue, trying to use my local inference llama.cpp server but OnlyOffice doesn't know what to do.

[Hello-World-Traveler]

@Splarkszter ollama works with this, where LocalAI doesn't work I have ollama to fill the gab.