Uncensored AI Chat and Roleplay API

Also known as: Uncensored Chat · NSFW AI chat API · AI roleplay chat

Uncensored AI Chat is Vellria's conversational model: you write, it answers, and it stays in whatever role you hand it. It is made for roleplay, collaborative fiction and open-ended adult conversation, and it is served raw, the same way in the playground on this page as through the API. If you came looking for an uncensored ChatGPT for stories and characters, this is the part of Vellria that does that job, paid from credits rather than a monthly subscription.

Roleplay runs through the system prompt. Describe the character, the setting and how the scene should feel, then talk to it as the user; the model keeps that voice from turn to turn as long as the earlier messages travel with each request. Nothing persists on our side between requests, so the conversation, the character sheet and any memory you want stay in your own app. That is how it serves as a character AI alternative you control instead of a persona someone else hosts.

The API speaks the OpenAI chat format. Point an OpenAI SDK at the Vellria base URL, use a vll_ key from the console and the model id from the quickstart; streaming, usage counts and the usual message roles behave the way existing code expects. Every response also reports what it cost in credits and dollars, so an app can keep a running total without a second call.

It reads pictures as well as text. Upload an image, pass its address as an image_url part in a user message and ask about it: describe a character from a portrait, continue a scene from a frame, write a caption. When the picture itself is the goal, Seedream 5.0 Pro makes stills and Wan 3.0 Pro makes clips, on the same credits and the same keys.

Each reply is charged per token from your credits. Fractions of a credit add up on your account and are taken from the balance as whole credits.

Any OpenAI SDK works. Set the base URL to https://vellria.com/v1, use your vll_ key, and pass the model id vellria-uncensored-chat.

curl https://vellria.com/v1/chat/completions \
  -H "Authorization: Bearer vll_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "vellria-uncensored-chat",
    "messages": [
      {"role": "system", "content": "You are Mira, a sarcastic space pirate. Stay in character."},
      {"role": "user", "content": "Mira, where are we headed tonight?"}
    ]
  }'

from openai import OpenAI

client = OpenAI(api_key="vll_...", base_url="https://vellria.com/v1")

stream = client.chat.completions.create(
    model="vellria-uncensored-chat",
    messages=[
        {"role": "system", "content": "You are Mira, a sarcastic space pirate. Stay in character."},
        {"role": "user", "content": "Mira, where are we headed tonight?"},
    ],
    stream=True,
)
for chunk in stream:
    if chunk.choices:
        print(chunk.choices[0].delta.content or "", end="")

import OpenAI from "openai"

const client = new OpenAI({ apiKey: "vll_...", baseURL: "https://vellria.com/v1" })

const stream = await client.chat.completions.create({
  model: "vellria-uncensored-chat",
  messages: [
    { role: "system", content: "You are Mira, a sarcastic space pirate. Stay in character." },
    { role: "user", content: "Mira, where are we headed tonight?" },
  ],
  stream: true,
})
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "")
}

Every field, the streaming format and the error codes are in the chat section of the API docs. Keys are created on the API keys page.