Skip to content
In progress software Embedded & IoT Infrastructure

ForgeOS & ForgeControl

Governed knowledge-to-operations platform with per-station hardware control for verifiable execution and full traceability.

Last updated: February 7, 2026

Tech Stack

Python FastAPI PostgreSQL ESP32 MQTT Node-RED InfluxDB Grafana

The Problem

In most workshops and production environments, knowledge lives in people’s heads. Procedures live in binders nobody opens. And when something goes wrong, nobody can answer the simplest questions: why was this done? Based on what? Who approved it?

The gap between knowing something and doing it correctly — every time, verifiably — is where quality breaks down, incidents happen, and trust erodes.

What We’re Building

ForgeOS is the software layer. It takes raw knowledge — documents, specs, measurements, experience — and turns it into verifiable claims. Those claims become standard operating procedures. Every step traces back to its source. Nothing is assumed. Nothing is unverifiable.

ForgeControl is the hardware layer. It sits at each station as a physical control box: switching high-load equipment on and off, enforcing safety interlocks (E-stop, overtemperature, smoke detection), metering energy consumption, and logging every action. When ForgeOS says “follow this procedure,” ForgeControl makes sure the station is physically ready and records what actually happened.

Together, they close the loop:

Source → Claim → Procedure → Execution → Evidence

How They Connect

ForgeControl publishes telemetry, state changes, and events over MQTT using versioned data contracts. ForgeOS consumes those streams for governance, traceability, and orchestration. SOPs flow downstream from ForgeOS to ForgeControl, where they’re executed step by step with sensor-gated verification. Every step emits lifecycle events back upstream.

The result: a continuous, auditable chain from the document that informed a decision to the physical evidence that it was carried out.

ForgeOS — Key Capabilities

  • Knowledge ingestion — documents, files, feeds parsed with stable identifiers
  • Claim extraction — atomic, source-grounded statements with confidence scoring
  • Canonical normalization — deduplication, contradiction detection, provenance retention
  • SOP generation — procedures derived from claims, with explicit gap surfacing
  • Governance — state machines for review and approval with full audit trail
  • Execution evidence — recorded outcomes, not assumed compliance

ForgeControl — Key Capabilities

  • Per-station switching — 4 high-load outputs (contactor-driven) + 1 direct relay
  • Hard-wired safety — E-stop and overtemperature override all software commands
  • Energy metering — MID-certified Modbus meter per station
  • LAN-first architecture — ESP32 + Ethernet, no cloud dependency
  • Degraded mode — 24h local buffering when the network is unavailable, automatic sync on reconnect
  • SOP execution — schema-driven step-by-step procedures with sensor-gated verification

Current Status

Both projects are under active, governed development. ForgeOS has its core foundations in place (data models, identity system, storage layer, state machines, API scaffolding). ForgeControl is in Engineering Validation with its control-plane infrastructure running and firmware development underway.

Open Source

Both projects are developed in the open: