System Overview

Learn about the bizsupply platform capabilities and how to use it effectively.

---

What is bizsupply?

bizsupply is a cloud-native document processing platform that helps you:

• Ingest documents from various sources (Gmail, Outlook, IMAP, custom APIs)
• Classify documents automatically using AI (invoices, receipts, contracts, etc.)
• Extract structured data using customizable schemas (ontologies)
• Aggregate related documents and build relationships

The platform provides a flexible, plugin-based architecture where you can create custom document processing workflows.

---

Key Capabilities

• Multi-tenant Processing - Complete data isolation between users and organizations
• Dynamic Plugin System - Create custom plugins for any document processing task
• Ontology-driven Extraction - Define exactly what data to extract with configurable schemas
• External Source Integration - Connect Gmail, Outlook, IMAP, or custom APIs
• LLM-powered AI - Leverage Google Gemini and OpenAI for intelligent processing
• Secure Credentials - OAuth tokens and API keys stored securely
• RESTful API - Integrate with any application via standard HTTP APIs
• Real-time Updates - Server-Sent Events (SSE) for live job status

---

How It Works

Document Processing Flow

1. Ingest documents from your connected sources
   ↓
2. Documents are stored and parsed
   ↓
3. Classification plugins analyze and label documents
   ↓
4. Extraction plugins pull structured data based on your ontology
   ↓
5. Aggregation plugins link related documents
   ↓
6. Access your processed documents via API

The Plugin System

Plugins are built using the bizsupply-sdk package (pip install bizsupply-sdk), which provides base classes, data models, and a CLI for scaffolding and validation.

bizsupply uses four types of plugins, each with a specific role:

Plugins execute in order: Source → Classification → Extraction → Aggregation

---

Core Concepts

Documents

A document is any file you process through bizsupply. Each document has:

• Metadata - Source information, dates, sender, etc.
• Labels - Classification tags (e.g., "invoice", "receipt")
• Data - Extracted structured fields (e.g., total, date, vendor)
• Files - Original file and parsed text content

Pipelines

A pipeline is a configured workflow that combines:

• One or more plugins to execute
• One or more ontologies for extraction schemas
• Optional source filters to process specific sources

Ontologies

An ontology defines your extraction schema:

• Taxonomy - Hierarchical labels for classification
• Fields - What data to extract (name, type, required)

Jobs

When you execute a pipeline, a job is created to track:

• Execution status (pending, running, completed, failed)
• Documents processed
• Results and any errors

---

Authentication

All API requests require a JWT Bearer token:

Authorization: Bearer <your_jwt_token>

Tokens are obtained through OAuth2 login with Google or Microsoft.

---

Architecture Principles

Multi-Tenancy

Every resource is scoped to your user and tenant:

• All data queries automatically filter by your context
• File storage paths include your identifiers
• You can only access your own documents and plugins

Plugin Isolation

Plugins run in isolated environments:

• Each plugin has its own dependencies
• Plugin crashes don't affect other operations
• Plugins receive only the data they need

Scalability

The platform automatically scales:

• Handle large document volumes
• Process multiple jobs concurrently
• Scale up during peak usage
• Background tasks (auto-sync, data processing) run as Cloud Run worker pools for cost-efficient scaling

---

Next Steps

• Install the SDK → pip install bizsupply-sdk
• Learn core concepts → Key Concepts
• Build a plugin → Create a Plugin
• Process documents → Process Documents
• Define extraction schemas → Create an Ontology