Naagmani Logo
Naagmani
Protecting What Is Precious
Developers

Build on one API in minutes

OpenAI-compatible. Typed SDKs in six languages. A live playground and clear docs so you can go from install to production fast.

Quick start

From install to first request

Three steps: install the SDK, set your key, and send a request. That's it.

1Install

bash
$ npm install naagmani

2Set your key

.env
NAAGMANI_KEY=nm_live_••••••••••••

3Send a request

hello.ts
const res = await client.chat.create({
model: "auto",
messages: [{ role: "user", content: "Hi" }],
})
SDK examples

Typed SDKs for your stack

The same OpenAI-shaped interface across every language — including streaming, tool calls, and structured outputs.

chat.ts
import Naagmani from "naagmani"
const client = new Naagmani({ apiKey: process.env.NAAGMANI_KEY })
const stream = await client.chat.stream({
model: "auto", // smart route
messages: [{ role: "user", content: "Hello!"} }],
})
for await (const chunk of stream) process.stdout.write(chunk.delta)
API examples

A familiar REST surface

Standard OpenAI-compatible endpoints. Use any HTTP client or existing SDK — just change the base URL.

List models — GET /v1/models
GET /v1/models
Authorization: Bearer $NAAGMANI_KEY
// 200 OK
{
"data": [{ "id": "gpt-4.1" }, { "id": "claude-4-sonnet" }]
}
Embeddings — POST /v1/embeddings
POST /v1/embeddings
{
"model": "text-embedding-3-large",
"input": "Naagmani unifies every model."
}
// returns a 3072-dim vector
Playground

Try it right here

Pick a model, write a prompt, and watch Naagmani route and stream the response.

playground.naagmani.com
Press "Run" to send the request through Naagmani.

Ready to simplify AI infrastructure?

Start building today with a unified API, smart routing, and one dashboard for every model. Free to start — no credit card required.