⚡ Using API & Playground

Experience and test AI models online with the Playground, and make seamless API calls via OpenAI / Anthropic / Gemini standard-compatible endpoints.

Playground

The Playground provides a lightweight online debugging and testing environment, allowing you to engage in multi-turn conversations directly with models and verify results without writing any code.

Select Group and Model: Select the channel group and the specific model to test at the bottom right of the input box.
Configure Generation Parameters: Click the parameters button to adjust Temperature, Top P, Frequency Penalty, and Max Tokens as needed.
Enter Conversation Prompts: Type your test question or System Prompt in the bottom input box.
Send and Debug: Click "Send" to submit the request and view streaming responses, token consumption, and response latency in real time.

API URL & Authentication

The platform's standard-compatible Base URL is:

https://api.openrely.ai/v1

Pass your API key as a Bearer Token in the HTTP request header:

Authorization: Bearer sk-your-api-key

Taking Chat Completions as an example:

curl https://api.openrely.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-api-key" \
  -d '{"model":"gpt","messages":[{"role":"user","content": "Hello"}]}'

SDK Code Examples

Python (OpenAI SDK)

from openai import OpenAI

client = OpenAI(
    api_key="sk-your-api-key",
    base_url="https://api.openrely.ai/v1",
)

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

Claude Native Messages Protocol

curl https://api.openrely.ai/v1/messages \
  -H "x-api-key: sk-your-api-key" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "claude",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Gemini Native Format

curl "https://api.openrely.ai/v1beta/models/gemini:generateContent?key=sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{"contents": [{"parts": [{"text": "Hello"}]}]}'

Supported Endpoints Overview

The platform is fully compatible with mainstream LLM protocol specifications. The primary supported endpoints are as follows:

Category	Endpoint	Protocol & Description
Chat Completions	`POST /v1/chat/completions`	OpenAI Chat Completions compatible endpoint, supporting full streaming (SSE), Function Calling, and multimodal inputs.
Playground Chat	`POST /pg/chat/completions`	Dedicated endpoint for the console web playground, reusing unified routing, protocol conversion, billing, and retry pipelines.
Claude Messages	`POST /v1/messages`	Anthropic Claude native compatible endpoint, supporting seamless direct connection for official Claude SDKs and native clients.
Gemini Native Call	`POST /v1beta/models/{model}:{action}`	Google Gemini native format endpoint, supporting operations like `generateContent`, `streamGenerateContent`, etc.
Gemini Compatible Call	`POST /v1/models/{model}:{action}`	Handles Gemini-style model operation requests with a `/v1` prefix, automatically performing protocol parsing and conversion.
Embeddings	`POST /v1/embeddings`	Batch vectorization of text, automatically adapted to upstream Embeddings channels.
Rerank	`POST /v1/rerank`	Reranks candidate `documents` based on `query` relevance, suitable for RAG-enhanced retrieval.
Image Generation	`POST /v1/images/generations`	Text-to-image endpoint, supporting generative models such as DALL·E, FLUX, Midjourney, etc.
Image Editing	`POST /v1/images/edits`	Image editing and inpainting, supporting JSON and `multipart/form-data` upload requests.
Audio Transcription	`POST /v1/audio/transcriptions`	Whisper speech recognition, transcribing audio to text.
Text to Speech	`POST /v1/audio/speech`	TTS speech synthesis, rendering text into high-quality audio streams.
Moderation	`POST /v1/moderations`	Sensitive content filtering and compliance detection.
Responses API	`POST /v1/responses`	OpenAI Responses format endpoint, supporting context compression and advanced event handling.
Realtime	`GET /v1/realtime` (WebSocket)	OpenAI Realtime-style full-duplex, low-latency voice/text streaming endpoint.
Model List	`GET /v1/models`	Lists currently available models, supporting adaptive response formats based on OpenAI, Anthropic, and Gemini request headers.
Model Details	`GET /v1/models/{model}`	Retrieves configuration and capability details for a specific model.