Set up RisuAI with an uncensored OpenAI-compatible endpoint
RisuAI is a character chat app with a web version, desktop builds and a self-hosted server. Its Custom API model accepts any OpenAI-compatible URL, and the web version sends requests through RisuAI's own proxy, so the Wild West API works on all of them.
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.
- Open Settings, then Bot Settings.
- Set Model to Custom API.
- Set URL to https://wildwestapi.com/v1/chat/completions.
- Paste your sk-ww- key into Key/Password.
- Type outlaw-1 into Request Model.
- Leave Format on OpenAI Compatible and start a chat.
Does it work with the Wild West API?
Yes. RisuAI's fetch code sends requests from the web version through RisuAI's proxy server, not straight from the page, so the missing CORS preflight on the API does not block it. The desktop build makes the request natively, and a self-hosted RisuAI server makes it from the server. All three work.
The one setting that breaks this is plain fetch in the advanced settings. With it on, the web version calls the API straight from the browser and gets a network error. Leave it off.
About the URL
The Custom API URL field sends the request exactly where you point it. RisuAI also has an Autofill Request URL option, on by default, that adds /chat/completions when the URL ends in v1. Entering the full path is the safest choice:
https://wildwestapi.com/v1/chat/completions
The settings that matter for long roleplay
- Max Context Size controls how much chat history and lorebook text is sent. The models here accept 512K to 1M tokens, but cost scales with what you send each turn. 32K to 64K is a good working range; see context window.
- Max Response Size is the reply cap, the same idea as max tokens. 500 to 1000 suits roleplay.
- Turn on Streaming to see replies as they are written; see streaming.
- Because the request goes out as OpenAI chat format, your prompt template's system prompt becomes the system message.
Troubleshooting RisuAI
- 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: turn off the plain fetch option in advanced settings so the web version uses RisuAI's proxy again, or use the desktop build.
- Wrong model or empty reply: make sure Request Model holds the exact id, such as
outlaw-1, and Format is OpenAI Compatible.
FAQ
Is my key sent through RisuAI's server on the web version?
Yes. On the web version requests go through RisuAI's proxy, which is what lets it reach APIs without CORS. If you would rather not, use the desktop build, which calls the API directly.
Can I use the Anthropic format instead?
Format has an Anthropic Claude option, but the OpenAI Compatible format with the /v1/chat/completions URL is the tested route for this API.
What does Autofill Request URL do?
When on, it completes a URL that ends in v1 with /chat/completions. Typing the full path makes the setting irrelevant.