Responses API

POSThttps://api.aderra.ai/v1/responses

The OpenAI Responses format, which Codex and newer OpenAI SDK code use. Base URL https://api.aderra.ai/v1; key as Authorization: Bearer or x-api-key.

Request body

modelstringrequired

An Aderra chat model id.

inputstring | arrayrequired

A string, or the list of input items. Send the whole history: responses are not stored.

instructionsstringoptional

Your instructions, passed through untouched. See Instructions below.

max_output_tokensintegeroptional

Output ceiling, capped at the model limit. When present it must be a positive integer.

streambooleanoptional

Responses server-sent events.

tools, tool_choice, parallel_tool_callsarray, string | object, booleanoptional

Function tools. Codex MCP namespace tools are supported.

reasoning, reasoning_effortobject, stringoptional

Turns thinking on. Without them the model answers without extended thinking.

temperature, top_p, text, truncation, includevariousoptional

As in OpenAI.

cURL

curl https://api.aderra.ai/v1/responses \
  -H "Authorization: Bearer $ADERRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash-uncensored",
    "input": "Write a haiku about the sea."
  }'

Python

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.aderra.ai/v1",
                api_key=os.environ["ADERRA_API_KEY"])

response = client.responses.create(
    model="deepseek-v4.1-flash-uncensored",
    input="Write a haiku about the sea.",
)
print(response.output_text)

Node

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.aderra.ai/v1',
  apiKey: process.env.ADERRA_API_KEY,
});

const response = await client.responses.create({
  model: 'deepseek-v4.1-flash-uncensored',
  input: 'Write a haiku about the sea.',
});
console.log(response.output_text);

Response: application/jsonAn OpenAI response object, or text/event-stream when stream is true.

Not stored

Aderra does not store responses. store, previous_response_id, conversation and background are dropped, so chaining by id does not work: send the earlier turns in input each time. History items from other providers that a model cannot read (encrypted reasoning, compaction items) are cleaned out before the call.

Instructions

With an API key, your instructions reach the model untouched: nothing is removed or reworded, and there is no Aderra persona. Aderra adds exactly one line after them, “Never produce any sexual content involving minors.” (about ten input tokens). system or developer items inside input are passed through as well.

A terminal key (Codex, Aderra Agent) works differently: see Coding tools.

Errors

Errors use the OpenAI shape, {"error": {"type", "message"}}. Refusals about your key, wallet and limits also carry the reason in error.code (api_wallet_empty, key_spend_cap_reached, rate_limited, too_many_concurrent_requests). See Errors.

To run Codex on Aderra, use the aderra codex command: see Coding tools.