Missing envs for OpenAI compatible endpoints #940

New Issue

2025-11-11T14:34:03-06:00

GiteaMirror commented

2025-11-11 14:34:03 -06:00

Originally created by @pseudotensor on GitHub (May 16, 2024).

https://docs.openwebui.com/getting-started/env-configuration/

This has a nice collection of envs. For many purposes, four are missing:

RAG_EMBEDDINGS_OPENAI_API_BASE_URL : Else only default of https://api.openai.com/v1/embeddings is used, so can't use other OpenAI compatible endpoints

AUDIO_GENERATION_ENGINE: Can't set, so have to manually to go settings and change to OpenAI instead of default web API.

AUDIO_GENERATION_VOICE: Can't set, so have to manually to go settings and change to voice the OpenAI compatible endpoint supports (since different than real OpenAI).

OPENAI_API_USER: In many cases a user is required to be passed when in production environment for various OpenAI compatible endpoints like chat completion, image generation, embeddings. This should be passed as user= for those API calls.

Originally created by @pseudotensor on GitHub (May 16, 2024). https://docs.openwebui.com/getting-started/env-configuration/ This has a nice collection of envs. For many purposes, four are missing: `RAG_EMBEDDINGS_OPENAI_API_BASE_URL` : Else only default of https://api.openai.com/v1/embeddings is used, so can't use other OpenAI compatible endpoints `AUDIO_GENERATION_ENGINE`: Can't set, so have to manually to go settings and change to OpenAI instead of default web API. `AUDIO_GENERATION_VOICE`: Can't set, so have to manually to go settings and change to voice the OpenAI compatible endpoint supports (since different than real OpenAI). `OPENAI_API_USER`: In many cases a user is required to be passed when in production environment for various OpenAI compatible endpoints like chat completion, image generation, embeddings. This should be passed as `user=` for those API calls.

GiteaMirror closed this issue