SDKs

Glytos is a plain REST API described by OpenAPI, so every language works. The most common starting points have hand-written SDKs - one for the browser and five for the server - plus a CLI for your terminal; everything else is either a generated client or a plain HTTP call, and both are first-class.

Web

Voice calls over WebRTC, and text chat, from a browser.

Node

Server-side TypeScript: agents, threads, calls, sessions, webhooks.

Python

The same surface, server-side, in Python.

PHP

Server-side PHP, with Laravel support.

.NET

Server-side C#, with ASP.NET Core support.

Go

Server-side Go, with zero dependencies.

CLI

Manage Glytos from your terminal or a CI job.

Which one

The web SDK is for one job: talking to an agent from a browser, by voice or as text. It never sees your API key.

The Node, Python, PHP, .NET and Go SDKs are for your backend. They cover the same surface, method for method, so pick the one your stack already speaks: threads for text conversations, calls and phoneNumbers for voice, and the rest of the platform either way. The CLI wraps the same API for your terminal and CI.

Only the web SDK belongs in a browser. The server SDKs take an API key, and an API key in front-end code is a key anyone can read.

Any other language

Two ways, both supported.

Generate a client

Point any OpenAPI generator at the spec:

openapi-generator-cli generate \
  -i https://api.glytos.com/api/v1/openapi.json \
  -g java \
  -o ./glytos-java

Just call it

The API is small enough that a generated client is often more machinery than you need. A bearer header and JSON is the whole protocol - see Authentication for the same call in eight languages.