The AI gateway
for agents

# Add the qynio gateway

Set up qynio so this project can call 700+ AI models through a single API.

## Step 1: Get an API key

Have the user sign up at https://api.qyn.io/login and create an API key. Set it as `QYN_API_KEY` in the environment.

## Step 2: Use the OpenAI SDK with qynio's compat endpoint

```ts
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.QYN_API_KEY,
  baseURL: "https://api.qyn.io/v1",
});

const completion = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-4-5",
  messages: [{ role: "user", content: "Hello" }],
});
```

## When to use what

- **Drop-in OpenAI SDK swap (above)** — fastest path, works with any existing OpenAI SDK code.
- **Agent SDK (Python or TypeScript)** — building autonomous agents with tools, MCP, span tracing, multi-agent composition. /docsagents
- **Structured outputs** — typed input/output with schemas. /docsbuild/task/calls
- **Indexes** — semantic storage for RAG. /docscapabilities/indexes
- **Agent CLI** — launch existing agents (Claude Code, Codex, OpenCode, ...) on qynio's gateway. https://qyn.io/agent-cli

## Critical rules

- Read `QYN_API_KEY` from env, never hardcode
- Pass models as `provider/model` strings (e.g. `anthropic/claude-opus-4-7`, `openai/gpt-5`)
- Use baseURL `https://api.qyn.io/v1`
- No markup on tokens. 0% fees on every crypto deposit
- 700+ models available — see /docscapabilities/models

Full docs (agent-readable): https://qyn.io/llms-full.txt