NewPoint any agent at UberLLM — one key, every model, list price or less.

Open WebUI

Open WebUI adds UberLLM as an OpenAI API connection.

Before you start

Setup (admin connection, for everyone on the server)

  1. Go to Settings → Admin → Connections.
  2. Under Manage OpenAI API Connections, click ➕ Add Connection.
  3. Fill in:
    • URL: https://api.uberllm.dev/v1
    • API Key: your ull_... key (leave the auth type on Bearer)
  4. Open Advanced (optional but recommended):
    • Prefix ID: uberllm. Models then show as uberllm.z-ai/glm-5.2, so they never clash with models from your other connections. Open WebUI removes the prefix before it sends the request.
    • Model IDs: add the models you want, one at a time with + (see below). With this list set, Open WebUI shows only these models and stops loading the full list.
    • Headers: {"X-Title": "Open WebUI"} tells UberLLM which app sent the request.
  5. Leave API Type on Chat Completions, the setting tested here. UberLLM also serves the Responses API.
  6. Click Verify Connection, then Save. Saving alone does not test anything.

UberLLM lists about 450 models, including image and video models that do not chat. Without a Model IDs list every one of them appears in the menu and each page load waits for the full list, so add an allow-list (or turn on Cache Base Model List).

Setup with Docker environment variables

docker run -d -p 3000:8080 \
  -e OPENAI_API_BASE_URLS="https://api.uberllm.dev/v1" \
  -e OPENAI_API_KEYS="ull_your-key" \
  -e OPENAI_API_CONFIGS='{"0":{"prefix_id":"uberllm","model_ids":["deepseek/deepseek-v4.1-flash","google/gemini-2.5-flash","z-ai/glm-5.2"]}}' \
  -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main

These variables are read on the first start only. After that the value saved in Open WebUI's database wins, unless you set ENABLE_PERSISTENT_CONFIG=False. If you already run Open WebUI, use the admin screen instead.

Direct connection (each user brings their own key)

  1. The admin turns on Direct Connections under Settings → Admin → Connections (or sets ENABLE_DIRECT_CONNECTIONS=true).
  2. Each user opens User Settings → Connections, clicks +, and enters Base URL https://api.uberllm.dev/v1 and their own API Key.

With a direct connection the user's browser calls UberLLM itself and the key stays in that browser. UberLLM allows these browser calls.

Recommended models

Checked on 2026-10-30 through UberLLM.

Model id Context Notes
deepseek/deepseek-v4.1-flash 1,048,576 Fast; reasons before it answers
google/gemini-2.5-flash 1,048,576 Reads images
qwen/qwen3-vl-235b-a22b-instruct 256,000 Reads images, answers without a reasoning step

Common errors

What you see Cause and fix
Verify Connection fails with 401 Key missing or mistyped.
No models appear The URL has a trailing slash or a typo. It must be exactly https://api.uberllm.dev/v1.
The model menu or page loads slowly No Model IDs allow-list: see above.
A model from another connection disappeared Two connections returned the same id with no prefix, and Open WebUI kept one. Set Prefix ID.
402 insufficient_credits Balance empty. Add credit under Dashboard → Credits.