Introduction

Run open-weight models through OpenAI and Anthropic compatible inference endpoints.

Prism is an inference API for coding agents. Use the OpenAI client already in your application, change its base URL, and select a Prism model.

OpenAI base URL:   https://api.prisminference.com/v1
Anthropic base URL: https://api.prisminference.com

Prism serves DeepSeek-V4.1-Flash, DeepSeek-V4-Flash, GLM-5.3, Kimi K3, and Qwen. The same API key works across the supported wire formats.

API formats

Endpoint	Format	Purpose
POST /v1/chat/completions	OpenAI Chat Completions	Generate text, stream tokens, call tools, and return structured output.
POST /v1/responses	OpenAI Responses	Use Responses clients with stateless input items and typed events.
GET /v1/models	OpenAI Models	List models available to the authenticated API key.
GET /v1/models/{model}	OpenAI Models	Retrieve one available model by ID.
POST /v1/messages	Anthropic Messages	Use Prism models with Anthropic clients and Claude Code.
POST /v1/messages/count_tokens	Anthropic token counting	Count input tokens before creating a message.

Responses requests are stateless. Resend prior output items instead of using previous_response_id.

OpenAI-compatible request

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.PRISM_API_KEY,
  baseURL: "https://api.prisminference.com/v1",
});

const response = await client.chat.completions.create({
  model: "prism/glm-5.3",
  messages: [
    { role: "user", content: "Find the race condition in this worker: ..." },
  ],
});

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

Keep API keys on your server. Do not put them in browser code, client bundles, or public repositories.

Data handling

Prism applies strict zero data retention to API inputs and outputs. Request and response content is processed only to complete the request, is not persisted, and is never used for model training. Review the Privacy Policy and Terms of Service.

Start building

Create an API key

Create an account and issue a key from account settings.

Quickstart

Send a request with cURL, TypeScript, or Python.

Models

Compare model IDs, context windows, and intended workloads.

Chat Completions

Read the OpenAI-compatible request and response contract.

Responses

Use the stateless Responses API format and typed streaming events.

Anthropic Messages

Connect Anthropic SDKs and Claude Code.