Nscale

Nscale is an AI/GPU cloud that pairs serverless, OpenAI-compatible inference with on-demand GPU compute. The Serverless Inference API serves open models (Llama, Qwen, DeepSeek, GPT OSS, Mistral, Flux) at https://inference.api.nscale.com/v1 with pay-per-token billing, while the platform API provisions GPU clusters, compute instances, networks, and storage.

6 APIs 0 Features

AIGPUInferenceServerlessCloud Compute

APIs

Nscale Serverless Chat Completions API

OpenAI-compatible chat completions across Llama, Qwen, DeepSeek, GPT OSS, and Mistral models, with streaming (SSE), tool use, and multimodal vision inputs, served from Nscale se...

Nscale Serverless Completions API

OpenAI-compatible legacy text completions endpoint that generates a continuation for a supplied prompt against a selected serverless model.

Nscale Serverless Embeddings API

OpenAI-compatible embeddings endpoint returning vector representations of text for retrieval, clustering, and semantic search, served by models such as Qwen3 Embedding.

Nscale Serverless Image Generation API

OpenAI-compatible text-to-image endpoint creating images from prompts using diffusion models such as Flux.1 [schnell], billed per megapixel.

Nscale Models API

Lists serverless models available for inference with identifiers, context length, and per-token pricing metadata via the OpenAI-compatible models endpoint.

Nscale GPU Compute and Clusters API

Platform API for provisioning and managing GPU compute instances, networks, security groups, object storage, organizations, and projects on the Nscale AI cloud.

Collections

Nscale Serverless Inference API

OPEN

Pricing Plans

Nscale Plans Pricing

3 plans

PLANS

Rate Limits

Nscale Rate Limits

4 limits

RATE LIMITS

FinOps

Nscale Finops

FINOPS

Event Specifications

Nscale Chat Completions Streaming (HTTP + SSE)

AsyncAPI 2.6 description of Nscale's **chat completion streaming** surface. Nscale does not publish a WebSocket API. The only asynchronous / event-style transport documented for...

ASYNCAPI

Resources

Sources

opencollection: 1.0.0
info:
  name: Nscale Serverless Inference API
  version: '1.0'
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Chat
    type: folder
  items:
  - info:
      name: Create a chat completion
      type: http
    http:
      method: POST
      url: https://inference.api.nscale.com/v1/chat/completions
      body:
        type: json
        data: "{\n  \"model\": \"meta-llama/Llama-4-Scout-17B-16E-Instruct\",\n  \"messages\": [{ \"role\": \"user\", \"content\"\
          : \"Hello\" }],\n  \"stream\": false\n}"
    docs: Generates a chat completion response for the provided model and chat history. Set stream=true for an SSE token stream.
- info:
    name: Completions
    type: folder
  items:
  - info:
      name: Create a completion
      type: http
    http:
      method: POST
      url: https://inference.api.nscale.com/v1/completions
      body:
        type: json
        data: "{\n  \"model\": \"meta-llama/Llama-4-Scout-17B-16E-Instruct\",\n  \"prompt\": \"Once upon a time\",\n  \"max_tokens\"\
          : 128\n}"
    docs: Generates a completion response for the provided model and prompt.
- info:
    name: Embeddings
    type: folder
  items:
  - info:
      name: Create embeddings
      type: http
    http:
      method: POST
      url: https://inference.api.nscale.com/v1/embeddings
      body:
        type: json
        data: "{\n  \"model\": \"Qwen/Qwen3-Embedding-8B\",\n  \"input\": \"The quick brown fox\"\n}"
    docs: Returns a vector representation of a given input.
- info:
    name: Images
    type: folder
  items:
  - info:
      name: Create an image
      type: http
    http:
      method: POST
      url: https://inference.api.nscale.com/v1/images/generations
      body:
        type: json
        data: "{\n  \"model\": \"BlackForestLabs/FLUX.1-schnell\",\n  \"prompt\": \"A serene mountain lake at sunrise\",\n\
          \  \"n\": 1,\n  \"size\": \"1024x1024\"\n}"
    docs: Creates an image based on the provided text prompt.
- info:
    name: Models
    type: folder
  items:
  - info:
      name: List models
      type: http
    http:
      method: GET
      url: https://inference.api.nscale.com/v1/models
    docs: Returns a list of serverless models available for inference.
  - info:
      name: Retrieve a model
      type: http
    http:
      method: GET
      url: https://inference.api.nscale.com/v1/models/{model}
    docs: Returns metadata for a single model by identifier.