API Reference

One API key and one base URL for every OpenAI, Anthropic, Google, and xAI model — plus image, video, and audio generation — in the standard OpenAI-compatible format. If your app can call the OpenAI API, it can call 1K4: change two lines.

Quickstart

Base URL: https://api.1k4.ai/v1

Auth: send your project key as an OpenAI-style bearer token — Authorization: Bearer <your key>. Mint or copy a key on the install page.

curl

curl https://api.1k4.ai/v1/chat/completions \
  -H "Authorization: Bearer $ONEKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.2",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Python (openai)

from openai import OpenAI

client = OpenAI(
    base_url="https://api.1k4.ai/v1",
    api_key="YOUR_1K4_KEY",
)

resp = client.chat.completions.create(
    model="gpt-5.2",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Node (openai)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.1k4.ai/v1",
  apiKey: process.env.ONEKEY_API_KEY,
});

const resp = await client.chat.completions.create({
  model: "gpt-5.2",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(resp.choices[0].message.content);

The same key and base URL reach every provider — swap the model: request claude-sonnet-4-5 and it routes to Anthropic, gemini-2.5-pro to Google, grok-4-0709 to xAI. No extra accounts, no code changes.

Base URL & compatibility

The native primary base URL is https://api.1k4.ai/v1. For compatibility, the legacy hosts remain fully supported and route to the same engine: https://api.1key4ai.com/v1 and the apex https://1k4.ai/v1. 1K4 works with any OpenAI SDK (Python, Node, Go) and any app with a custom-OpenAI-endpoint setting — Cursor, TypingMind, Open WebUI, LibreChat, Continue.dev, and more.

Models & routing

Request a model by its name, or by a short alias (for example claude-sonnet-4-5 or gemini-flash-latest) and 1K4 routes it to the right provider automatically. Text models come from OpenAI, Anthropic, Google, and xAI; media adds image, video, and audio generation (OpenAI, Google, fal.ai for images; Sora for video; OpenAI for speech and transcription).

For the current, authoritative list see the pricing page or call GET /v1/models with your key.

Endpoints

/v1/chat/completionsChat + streaming (OpenAI format)
/v1/completionsLegacy text completion
/v1/messagesAnthropic Messages format
/v1/responsesResponses API (reasoning models)
/v1/modelsList the models your key can call
/v1/embeddingsText embeddings
/v1/images/generationsImage generation + /v1/images/edits
/v1/audio/speechText-to-speech, plus /v1/audio/transcriptions + /translations
/v1/videosVideo generation (async)

Also available: /v1/rerank, /v1/moderations, and a realtime websocket at /v1/realtime. Not supported: /v1/files, /v1/fine-tunes, and /v1/images/variations.

Good to know

Reasoning models omit temperature. Reasoning / Responses-pathway models — for example gpt-5-pro, gpt-5.4-pro, o3/o4, codex, and deep-research models — reject the temperature parameter. Omit it for those.
Media billing. Image, video, and audio models use their own endpoints (above) and are billed per unit, not per token.
Local models. Ollama / local models run through the 1K4 bridge on your own machine, not this cloud endpoint.
Billing protection. A missing or invalid key returns 401 with a JSON error object; requests stop cleanly at a zero balance (no debt).