Connect Janitor AI to llmcloud.ai in 5 minutes

Janitor AI lets you bring your own OpenAI-compatible endpoint. llmcloud.ai serves open models like DeepSeek, Llama 4, Kimi and Mistral from our own data centers, so long roleplay threads get fast, low-cost tokens without you thinking about capacity.

Step 1 — Create an API key

Sign in to llmcloud.ai, open the console and create a new key.

Step 2 — Open Janitor AI proxy settings

In Janitor AI, open your profile, then API Settings, then Using Proxy, and choose Custom (OpenAI compatible).

Step 3 — Configure the endpoint

Base URL:  https://api.llmcloud.ai/v1
API key:   <your llmcloud key>
Model:     deepseek-v3.1

Click Check API Key / Model to confirm the connection.

Step 4 — Pick a model

• deepseek-v3.1 — long, coherent roleplay
• llama-4-maverick — fast and creative with personas
• kimi-k2 — very long context for long threads
• mistral-large-3 — tight prose

Troubleshooting

• Invalid API key: copy the whole key with no trailing space.
• Model not found: use a model ID from the catalog at llmcloud.ai/models.
• Slow first reply: the first request on a cold thread can take a moment; later replies stream immediately.