Skip to content

02-platform-overview.md

Latest commit

 

History

History
121 lines (93 loc) · 4.34 KB

02-platform-overview.md

File metadata and controls

121 lines (93 loc) · 4.34 KB
title Platform Overview
category platform
order 2
description What Busibox is and what it can do for you
published true

Platform Overview

Busibox is a self-hosted AI platform that gives your organization document processing, semantic search, AI agents, and custom applications — while keeping your data private and under your control.

What Busibox Does

Busibox combines several capabilities into one platform:

  • Document processing — Upload PDFs, Word files, spreadsheets, images, and more. The system extracts text, chunks it, and makes it searchable.
  • Semantic search — Ask questions in natural language and get answers grounded in your documents. No need to remember exact keywords.
  • AI agents — Chat with assistants that can search your documents, browse the web, and help with tasks.
  • Custom apps — Specialized tools like Status Report and Recruiter run on the same platform, with the same security and data access.

Why It Matters

Most AI platforms force a choice: use a powerful cloud service and send your data elsewhere, or run limited tools locally. Busibox removes that trade-off.

  • Your data stays private. Everything runs on your organization's infrastructure. Documents, conversations, and search indexes stay on your network.
  • Local and cloud AI. Your admin can configure local models for speed and privacy, cloud models for complex tasks, or a mix of both.
  • Security by design. You only see documents you're allowed to see. Agents respect the same permissions — they can't access data you can't access.

Core Capabilities

Capability What it means for you
Document processing Upload almost any file format. The system handles extraction, chunking, and indexing automatically.
Semantic search Ask questions in plain language. The system finds relevant passages even when you don't use the exact words from the document.
AI agents Chat with assistants that search your documents, browse the web, and answer questions with citations.
Custom apps Use specialized apps (project tracking, cost estimation, etc.) that share your authentication and document access.

How the Platform Fits Together

The diagram below shows how the main pieces connect. As a user, you interact with the apps (AI Portal, Agent Manager, etc.). Those apps talk to AI services, data services, and security — all behind the scenes.

graph TB
    subgraph users [Users]
        Browser[Browser]
    end

    subgraph proxy [Reverse Proxy]
        Nginx[nginx]
    end

    subgraph apps [Applications]
        Portal[AI Portal]
        AgentMgr[Agent Manager]
        UserApps[Custom Apps]
    end

    subgraph ai [AI Services]
        AgentAPI[Agent API]
        LiteLLM[LiteLLM Gateway]
        vLLM[vLLM / MLX]
        Frontier[OpenAI / Bedrock]
    end

    subgraph data [Data Services]
        DataAPI[Data API]
        Worker[Data Worker]
        SearchAPI[Search API]
        EmbedAPI[Embedding API]
    end

    subgraph infra [Infrastructure]
        PG[(PostgreSQL)]
        Milvus[(Milvus Vectors)]
        MinIO[(MinIO Storage)]
        Redis[(Redis Queue)]
    end

    subgraph security [Security]
        AuthZ[AuthZ Service]
    end

    Browser --> Nginx
    Nginx --> Portal
    Nginx --> AgentMgr
    Nginx --> UserApps

    Portal --> AuthZ
    Portal --> DataAPI
    Portal --> AgentAPI
    AgentMgr --> AgentAPI

    AgentAPI --> LiteLLM
    AgentAPI --> SearchAPI
    LiteLLM --> vLLM
    LiteLLM --> Frontier

    DataAPI --> PG
    DataAPI --> MinIO
    Worker --> PG
    Worker --> MinIO
    Worker --> Milvus
    Worker --> EmbedAPI
    SearchAPI --> Milvus
    SearchAPI --> PG
    EmbedAPI --> LiteLLM

    AuthZ --> PG
Loading

Typical User Flow

A typical session looks like this:

  1. Authenticate — Log in to the AI Portal with passkey, TOTP, or magic link.
  2. Upload — Add documents. They're stored, processed, and indexed automatically.
  3. Search or chat — Ask questions. The system finds relevant passages and, when you use an agent, synthesizes answers with citations.
  4. Use apps — Open Agent Manager, Document Manager, Chat, or other apps. They use the same documents and permissions.

Everything is connected: your documents power search and agents, and your permissions control what you can see across the platform.