OpenAI-compatible API

What you need

1. An API key with the completions scope. See API keys and scopes.

2. The base URL https://idapt.app/api/openai/v1.

3. An active subscription, or your own provider key (BYOK). The free daily pool does not apply on this gateway.

Quick start

from openai import OpenAI

client = OpenAI(
    base_url="https://idapt.app/api/openai/v1",
    api_key="uk_your_key_here",
)

response = client.chat.completions.create(
    model="openai/gpt-5.4-pro",
    messages=[{"role": "user", "content": "Hello!"}],
)

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

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://idapt.app/api/openai/v1",
  apiKey: "uk_your_key_here",
});

const response = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-4.6",
  messages: [{ role: "user", content: "Hello!" }],
});

console.log(response.choices[0].message.content);

curl https://idapt.app/api/openai/v1/chat/completions \
  -H "Authorization: Bearer uk_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-flash",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Supported endpoints

• POST /chat/completions: streaming and non-streaming chat.

• POST /responses: a non-streaming bridge. Sending stream: true returns 501, so use chat/completions for streaming.

• POST /embeddings: text embeddings.

• POST /images/generations: image generation.

• POST /audio/speech and POST /audio/transcriptions: text to speech and speech to text.

• POST /videos: the async video family (submit, then poll).

• GET /models and GET /models/{id}: the model catalog.

What is rejected

A few OpenAI parameters have no idapt equivalent, and idapt fails them loudly rather than pretending. A request with n > 1, or with the special auto model, is rejected with a clear error so your app cannot drift without you noticing.

Streaming

Set stream: true on chat/completions to receive server-sent events, which every OpenAI SDK understands. Add stream_options: { include_usage: true }to get token counts in the final chunk.

Usage and billing

Completions on this gateway need an active subscription (included allowance first, then Usage Credits as overflow) or your own provider key. Calls without one return a 403 with code free_tier_requires_subscription. Language models bill at the prices shown in the app, with no separate API pricing; bring-your-own-key calls bill to your provider account and draw no idapt credits. Completions are limited to 60 requests per minute.

Watch your spend on the usage page (/app/usage), and read Plans, allowances, and credits for how the meters fit together.