PLATFORM
SDKs & libraries
Official server-side SDKs wrap the REST API in idiomatic code, handle auth and retries, and ship typed models for every resource. They're thin on purpose — anything you can do in an SDK maps directly to an endpoint in these docs.
Server-side only. SDKs use your secret key and must run on a server you control. Never ship a secret key in a browser or mobile app; for client apps, mint short-lived tokens from your backend.
Install
| Language | Install |
|---|---|
| Node.js | npm install atommatrix |
| Python | pip install atommatrix |
| PHP | composer require atommatrix/atommatrix-php |
| Go | go get github.com/atommatrix/atommatrix-go |
| Java | implementation "ai.atommatrix:atommatrix:1.x" |
| Ruby | gem install atommatrix |
| .NET | dotnet add package AtomMatrix |
Quick examples
Node.js
import AtomMatrix from "atommatrix";
const atom = new AtomMatrix(process.env.ATOMMATRIX_API_KEY);
const message = await atom.messages.send({
to: "+14155550123",
from: "ATOMMTX",
body: "Your code is 481920"
});
console.log(message.id, message.status);
Python
import os
from atommatrix import AtomMatrix
atom = AtomMatrix(os.environ["ATOMMATRIX_API_KEY"])
verification = atom.verifications.start(to="+14155550123", channel="sms", fallback="voice")
print(verification.status) # "pending"
Go
client := atommatrix.New(os.Getenv("ATOMMATRIX_API_KEY"))
call, err := client.Calls.Create(ctx, &atommatrix.CallParams{
To: "+14155550123",
From: "+14155550999",
Flow: []atommatrix.Action{{Action: "say", Text: "Hello from AtomMatrix."}},
})
Configuration
- Read the key from an environment variable (
ATOMMATRIX_API_KEY); don't hard-code it. - SDKs retry idempotent failures (
429/5xx) with backoff and attach an idempotency key automatically on writes. - Each SDK exposes the raw response and
request_idfor logging. - Pin a major version and read the changelog before upgrading across majors.
Help & community
SDKs are open source on GitHub — issues and pull requests welcome. For account-specific questions or to report a security concern, email developers@atommatrix.ai. For a production incident, contact support with the request_id from the affected call.