Skip to main content
Let people speak in their own language and be understood in another, live. Glot gives developers the infrastructure to build real-time translated conversations into voice, video and in-person products. Stream microphone audio into a conversation and receive translated speech and structured transcripts as it unfolds, without building and operating the underlying speech, translation, voice and real-time media stack yourself. The platform combines a server-side REST API for creating and managing conversations with a WebRTC session for streaming audio and receiving translations.

​
Before you start

Base URL

Send API requests to https://api.staging.glot.com.

Authentication

Authenticate server-side requests using a Glot API key as a bearer token.

Prepaid credit

Your organisation needs a positive credit balance before creating a translation room.
Test your API key by requesting the supported languages:
Keep your Glot API key on your server. Never expose it in browser, mobile or other client-side code.

​
Core concepts

A translated conversation is built from three components.

Room

The translation session created and managed through the REST API.

Participant token

A short-lived credential that identifies a participant and declares their language.

Real-time session

The WebRTC connection that carries microphone audio, translated speech and transcript events.

​
How Glot works

1

Create a room

Call POST /v1/rooms from your server to create a translation session.
2

Issue participant tokens

Call POST /v1/rooms/{room_name}/token for each participant, providing their identity and language.
3

Connect the clients

Return the generated url and token to the corresponding client. The client uses them to connect to the room over WebRTC.
4

Publish microphone audio

Each connected client publishes its microphone audio to the room.
5

Receive the translation

Play the translated audio tracks and render the live transcript events returned through the real-time session.
6

Close and review

Close the room through the API to disconnect its participants and stop billing. You can then retrieve the final usage and charge.

​
Conversation modes

Glot supports separate-device and shared-device conversations.

Dual-device

Each participant joins from their own device and declares one language. Glot translates between the two participants in real time.

Shared-device

Two people share one device. The connection declares both languages and carries the conversation through a single audio stream.
The translation service automatically joins once two languages are present in the room.

​
Platform capabilities

Translation sessions

Create and manage rooms, configure participant languages, and receive translated speech and transcripts.

Secure access

Keep API keys on your backend and issue short-lived, room-specific credentials to clients.

Usage and billing

Retrieve room-level and organisation-level usage, participant minutes, charges and credit movements.

API-key management

Create, list and revoke API keys as your integration and team grow.

Invite links

Let someone without an account join a single room, for as long as you choose.

​
Start building

Quickstart

Create a room and stream your first translated conversation.

API and streaming

Understand how the REST API and real-time session work together.

API reference

Explore rooms, participant tokens, usage, billing and API-key endpoints.