Skip to content

introduction.mdx

Latest commit

 

History

History
132 lines (104 loc) · 4.66 KB

introduction.mdx

File metadata and controls

132 lines (104 loc) · 4.66 KB
Raw
title InteraOne
description Learn what InteraOne is, what it includes, and how its agentic customer-support architecture fits together.
sidebarTitle Introduction
icon house

InteraOne is an open-source, multi-tenant customer-support platform that combines a shared inbox, tickets, contacts, knowledge retrieval, an embeddable widget, messaging channels, and AI agents in one deployable system.

Our mission is to unify every customer communication channel with a single shared AI brain that businesses fully own and control. We empower organizations to build, self-host, and scale intelligent support without vendor lock-in.

Start the complete development stack with pnpm, Docker, and one Make target. Deploy InteraOne to Production with a One-Click Server Setup Follow requests across the gateway, queues, AI agent, worker, and data stores. Understand how a single shared brain is decoupled to handle inbound and outbound communication across all channels.

What the platform includes

Provider routing, streaming replies, RAG, tools, agent assist, and human handoff. Conversations, assignment, contacts, tickets, templates, notifications, and analytics. Web widget, email, WhatsApp, Telegram, QR entry points, and direct links.

System at a glance

flowchart LR
    %% =========================
    %% Inbound Channels
    %% =========================
    subgraph Channels["Communication Channels"]
        Web[Embeeded Agent]
        WhatsApp[WhatsApp]
        Telegram[Telegram]
        Email[Email]
    end

    %% =========================
    %% Gateway
    %% =========================
    Gateway[Gateway + WebSocket]

    Web --> Gateway
    WhatsApp --> Gateway
    Telegram --> Gateway
    Email --> Gateway
  

    %% =========================
    %% Infrastructure
    %% =========================
    Mongo[(MongoDB)]
    Redis[(Redis)]
    MinIO[(MinIO)]

    Gateway --- Mongo
    Gateway --- Redis
    Gateway --- MinIO

    %% =========================
    %% Processing Pipeline
    %% =========================
    Gateway --> InQueue[[Inbound Queue]]

    InQueue --> Agent[Shared AI Agent]

    Agent --> Webhook[Webhook Engine]

    Webhook --> Outbound[Outbound Delivery]

    %% =========================
    %% Outbound Channels
    %% =========================
    Outbound --> WebOut[Embedded Agent]
    Outbound --> WhatsAppOut[WhatsApp]
    Outbound --> TelegramOut[Telegram]
    Outbound --> EmailOut[Email]
Loading

The repository is a pnpm/Turborepo monorepo. Its five application packages have deliberately separate responsibilities:

Application Responsibility
apps/gateway Express REST API, Socket.IO, authorization, domain modules, storage, and queue producers/consumers
apps/console Vite + React operator dashboard for conversations, tickets, contacts, knowledge, channels, widgets, and settings
apps/launcher Iframe-isolated browser widget loader and widget UI bundle
apps/agent AI reply, ingestion, analysis, assist, tool execution, model routing, and vector retrieval
apps/worker Email delivery, analytics persistence, any background processing
The source repository contains an open-core `ee/` area. Features loaded from it are governed by `ee/LICENSE` and runtime entitlement checks; review that license before commercial use.

Choose your path

Follow the [quickstart](/quickstart), sign up through the local console, configure a model provider, add knowledge, and embed the widget on a test page. Start with [self-hosting architecture](/self-hosting/architecture), then configure environment variables, Caddy, persistent volumes, backups, and monitoring. Read the [contributor overview](/contributing/overview), run the focused checks for the package you change, and open a scoped pull request. Search understands page titles, headings, descriptions, and code. Press ⌘ K or Ctrl K to find a feature quickly.