Connect Open WebUI to an uncensored OpenAI-compatible API
Open WebUI is a self-hosted chat interface with its own backend, so it can call the Wild West API server-side with no browser restrictions. You add it once as an OpenAI connection in the admin settings.
Set it up
- Sign in at wildwestapi.com/signup/ with the six-digit code sent to your email, open the dashboard and create an API key. It starts with sk-ww- and is shown in full once, so copy it somewhere safe.
- In Open WebUI, open Settings, then Admin Settings, then Connections.
- Under the OpenAI API section (Manage OpenAI API Connections), click the plus button to Add Connection.
- Set URL to https://wildwestapi.com/v1 and paste your sk-ww- key into API Key.
- Leave the connection type on the default and save. Open WebUI reads the model list from /v1/models.
- If the models do not appear, type outlaw-1 into Model IDs (Filter), click the plus icon, and save again.
- Start a new chat and pick outlaw-1 from the model selector at the top.
Does it work with the Wild West API?
Yes. A normal Open WebUI connection is called by the Open WebUI server, not by your browser tab, so the fact that the Wild West API does not answer CORS preflight does not matter. Docker and pip installs behave the same way.
The exception is Direct Connections, where the browser talks to the provider itself. Open WebUI's docs say that mode needs CORS on the provider, so it fails today. Use an admin connection. Pick outlaw-1 to start, or another id from the models page.
The settings that matter for long roleplay
- Max Tokens in a model's advanced parameters sets the reply length. Without it the model decides, and long scenes can end mid sentence. 600 to 1200 is a sensible range for prose; see max tokens.
- Context. Open WebUI sends the conversation history with each request. These models take 512K tokens or more of context window, so long chats fit, but you pay for every input token each turn.
- Stream Chat Response should stay on so text arrives as it is written. See streaming.
- System Prompt on a workspace model is the place for your character card or narrator rules; see system prompt.
- Temperature around 0.8 to 1.0 suits creative writing.
Troubleshooting Open WebUI
- 401: the key is missing, mistyped or disabled. Paste it again with no trailing space and check it starts with
sk-ww-. A key from another provider will never work here. - 402: your balance is too low for the request. Top up in the dashboard; see pricing. A huge context on every turn burns credit faster than you might expect.
- 404 or 400 with an unknown model: the model id is wrong. Use an exact id from the model list, such as
outlaw-1, with no provider prefix and no spaces. - 429: you are rate limited. Wait a few seconds and send again, and turn off any auto-retry loop that fires instantly.
- 5xx: trouble upstream. Retry; if it keeps happening, switch to another model id for a while.
- Network error, "Failed to fetch" or a CORS message: you are probably using a Direct Connection from user settings. Remove it and add the API under Admin Settings, Connections instead, so the server makes the call.
- Verify Connection fails but chat works: Open WebUI's docs note that the verify button only checks the /models endpoint. If chatting works, the connection is fine.
FAQ
Do I need to run Open WebUI with Docker for this to work?
No. Any install that runs the Open WebUI backend works, because the backend makes the API call. Docker, pip and the desktop app are all fine.
Can each user bring their own Wild West key?
Not through Direct Connections today, because those run in the browser and the API does not answer CORS preflight. An admin can add the connection once, or add several connections with different keys.
Why does a chat get more expensive as it goes on?
The whole history is resent each turn, so input tokens grow. Start a new chat to reset.