POST/responsesUpdated

Responses API for abliterated models

POST /responses is OpenAI's newer format, and here it is stateless. Send the whole conversation in input with store set to false.

https://api.silkcompute.com
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.silkcompute.com",
    api_key=os.environ["SILK_API_KEY"],
)

response = client.responses.create(
    model="deepseek-v4.1-flash-1017-uncensored",
    input="What is a mutex?",
    store=False,
)
print(response.output_text)

Example response

{
  "object": "response",
  "model": "deepseek-v4.1-flash-1017-uncensored",
  "output": [
    {"type": "reasoning"},
    {
      "type": "message",
      "role": "assistant",
      "content": [
        {"type": "output_text", "text": "A lock only one thread can hold at a time."}
      ]
    }
  ],
  "usage": {
    "input_tokens": 16,
    "output_tokens": 48,
    "input_tokens_details": {"cached_tokens": 0}
  }
}
import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://api.silkcompute.com',
  apiKey: process.env.SILK_API_KEY,
})

const response = await client.responses.create({
  model: 'deepseek-v4.1-flash-1017-uncensored',
  input: 'What is a mutex?',
  store: false,
})
console.log(response.output_text)

Example response

{
  "object": "response",
  "model": "deepseek-v4.1-flash-1017-uncensored",
  "output": [
    {"type": "reasoning"},
    {
      "type": "message",
      "role": "assistant",
      "content": [
        {"type": "output_text", "text": "A lock only one thread can hold at a time."}
      ]
    }
  ],
  "usage": {
    "input_tokens": 16,
    "output_tokens": 48,
    "input_tokens_details": {"cached_tokens": 0}
  }
}
curl https://api.silkcompute.com/responses \
  -H "Authorization: Bearer $SILK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash-1017-uncensored",
    "input": "What is a mutex?",
    "store": false
  }'

Example response

{
  "object": "response",
  "model": "deepseek-v4.1-flash-1017-uncensored",
  "output": [
    {"type": "reasoning"},
    {
      "type": "message",
      "role": "assistant",
      "content": [
        {"type": "output_text", "text": "A lock only one thread can hold at a time."}
      ]
    }
  ],
  "usage": {
    "input_tokens": 16,
    "output_tokens": 48,
    "input_tokens_details": {"cached_tokens": 0}
  }
}
response = client.responses.create(
    model="deepseek-v4.1-flash-1017-uncensored",
    store=False,
    input=[
        {"role": "user", "content": "Name a lock-free queue."},
        {"role": "assistant", "content": "A Michael-Scott queue."},
        {"role": "user", "content": "How does it handle the ABA problem?"},
    ],
)
const response = await client.responses.create({
  model: 'deepseek-v4.1-flash-1017-uncensored',
  store: false,
  input: [
    { role: 'user', content: 'Name a lock-free queue.' },
    { role: 'assistant', content: 'A Michael-Scott queue.' },
    { role: 'user', content: 'How does it handle the ABA problem?' },
  ],
})
curl https://api.silkcompute.com/responses \
  -H "Authorization: Bearer $SILK_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash-1017-uncensored",
    "store": false,
    "input": [
      {"role": "user", "content": "Name a lock-free queue."},
      {"role": "assistant", "content": "A Michael-Scott queue."},
      {"role": "user", "content": "How does it handle the ABA problem?"}
    ]
  }'

How it works

StatelessSend every turnNothing is kept between calls, so each request carries every earlier turn. Keep them unchanged to get cache hits.ThinkingA reasoning itemThinking comes back as a reasoning output item, and reasoning.effort sets it.
Not supportedThree things to skipprevious_response_id is not supported, video is dropped and still billed, and hosted tools such as web search are not available.

Parameters

ParameterTypeRequiredWhat it does
modelstringA model id from the model list. More
inputstring or arrayThe whole conversation, every time. Nothing is kept between calls.
storebooleanSet it to false. The endpoint stores nothing.
reasoning.effortstringlow, high or max. More
toolsarrayFunction definitions. Hosted tools such as web search are not available. More
streambooleanStreams the reply as events. More