Inference API

The ParalonCloud Inference API lets you call open large language models over HTTP — no GPU to provision, no model to download. It's OpenAI-compatible, so any tool or SDK that speaks the OpenAI API works by changing two lines: the base URL and the key. Requests are routed to GPU nodes across the network that can serve the model you ask for.

Base URL

https://paraloncloud.com/v1

The API mirrors the OpenAI surface — /v1/chat/completions, /v1/responses, /v1/embeddings, /v1/classify, /v1/images/generations, /v1/batches, /v1/models — and authenticates with a bearer key. It also answers Anthropic's Messages format at /v1/messages, which is what Claude Code speaks.

Quick start

Point the official OpenAI Python SDK at ParalonCloud:

from openai import OpenAI

client = OpenAI(
    api_key="prlc_your_key_here",
    base_url="https://paraloncloud.com/v1",
)

response = client.chat.completions.create(
    model="qwen3.6-27b",
    messages=[{"role": "user", "content": "Hello!"}],
)

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

Or with curl:

curl https://paraloncloud.com/v1/chat/completions \
  -H "Authorization: Bearer prlc_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.6-27b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

What's next

• Authentication — create an API key and make your first authenticated call.
• Chat completions — the request and response format, parameters, and streaming.
• Models — list available models and pick the right one.
• Embeddings — turn text into vectors for search and RAG, billed per input token.
• Classify — a label and calibrated probabilities per input, no generated text: rerank, grade, guard, route. Billed per input token.
• Responses API — the OpenAI Responses shape (input items, function tools, streaming events) on the same models.
• Messages API — the Anthropic Messages shape (system blocks, tool_use and tool_result, streaming events), so Claude Code and the Anthropic SDKs work by changing the base URL.
• Batch API — a JSONL file of requests in, a file of results out within 24 hours, at half the token price.

Need GPUs too?

The same prlc_ key can also rent GPUs programmatically — browse available hardware, start a rental, and stop it over HTTP. One key and one balance cover both APIs, so you can build agents that call models and rent the GPUs to run them.

Try it without code

Want to test a model before you write any code? Open the Playground to chat with the available models in your browser, or head to the Console to create keys and watch your usage.

A free key carries a 250,000-token trial, unlocked in the Console with GitHub, Discord or a post on X (see Credits & Billing); after that a premium key pays per token from prepaid credits. See Credits & Billing for prices and limits.