API reference | Uncensured AI

POST/v1/chat/completions

Generate a response

Send the conversation in messages and select the model by its exact ID. These examples use GLM 5.3 Flash.

Install the openai package and set UNCENSURED_API_KEY as shown in the quickstart.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://uncensuredai.com/v1",
    api_key=os.environ["UNCENSURED_API_KEY"],
)

response = client.chat.completions.create(
    model="glm-5.3-flash-abliterated",
    messages=[{"role": "user", "content": "Scrivi un breve saluto in italiano."}],
    reasoning_effort="low",
    max_tokens=4096,
)

print(response.choices[0].message.content)

The fields in the example

Field	How to use it
`model`	An exact model ID returned by GET /v1/models for your key.
`messages`	A list of messages, each with role and content.
`max_tokens`	The output budget, which can include reasoning tokens. The example uses 4096.
`reasoning_effort`	A level supported by the selected model. GLM Flash uses low in this example.
`stream`	Set to true to receive the response in successive chunks.

OpenAI compatibility refers to Chat Completions. Images, tools, context limits and reasoning options depend on the model and client.

Read the response

The generated text is in choices[0].message.content. Check finish_reason to understand how generation ended and usage for the reported token counts.

print(response.choices[0].message.content)
print(response.choices[0].finish_reason)
print(response.usage)

POST/v1/chat/completions

Display text as it arrives

Reuse the client from the Python example and enable streaming. Each event may contain only part of the response.

stream = client.chat.completions.create(
    model="glm-5.3-flash-abliterated",
    messages=[{"role": "user", "content": "Ciao!"}],
    reasoning_effort="low",
    max_tokens=4096,
    stream=True,
)

for chunk in stream:
    if chunk.choices:
        text = chunk.choices[0].delta.content
        if text:
            print(text, end="", flush=True)
print()

Some events have no text. Check choices and delta.content before displaying a chunk.

GET/v1/models

Find the models available to your key

Reuse the Python client to list model IDs. This checks the connection without generating a reply.

models = client.models.list()
for model in models.data:
    print(model.id)

Availability comes from your authenticated model list. The public capabilities catalogue separately describes confirmed limits and reasoning settings.

Open the model capabilities catalogue