Pharos Overview

What Pharos is, how it works, and why it matters for your manufacturing operations.

What is Pharos?

Pharos is an MCP (Model Context Protocol) server that gives AI agents live, structured access to your manufacturing ERP data. It connects to your existing databases (SYSPRO, Tempo MRP, analytics warehouses) and exposes them as tools that any MCP-compatible AI client can use.

Architecture

┌─────────────┐     ┌──────────────┐     ┌─────────────────┐
│  AI Client   │────▶│   Pharos     │────▶│  SYSPRO DB      │
│  (Claude,    │ MCP │   MCP Server │     │  Tempo MRP      │
│   GPT, etc.) │◀────│              │◀────│  Analytics WH   │
└─────────────┘     └──────────────┘     └─────────────────┘

Pharos sits between the AI client and your data. It translates natural-language tool calls into database queries and returns structured results.

Key Concepts

MCP Tools are functions the AI can call. Pharos exposes 50+ tools for querying, analysing, and writing to your ERP. Each tool has a defined schema, description, and parameters.

Stateless. No persistent daemon. Each session initialises fresh, connects to the databases, and terminates when done. No background processes, no state to manage.

curl-based. All communication happens over HTTP. No Python runtime, no special client library. If you can run curl, you can use Pharos.

Read + Write. Read operations (queries, reports, analysis) are unrestricted. Write operations (labour posting, inventory adjustments) go through a confirmation gate: the AI proposes, the human confirms.

Data Sources

Pharos connects to multiple databases simultaneously:

| Source | Purpose | Example Data | |--------|---------|-------------| | SYSPRO Company DB | Transactional data | Customers, orders, inventory, invoices | | SYSPRO Admin | Schema metadata | Table definitions, column types, relationships | | Tempo MRP | Planning data | MRP runs, shortages, suggestions | | Analytics Warehouse | Reporting | KPIs, cross-entity aggregations |

Who Uses Pharos?

Operations managers query MRP status, delivery schedules, production KPIs through natural language
Finance teams check customer risk, overdue invoices, credit limits
Production supervisors post labour, review work order progress
IT/developers use schema discovery, data model exploration, integration testing