An introduction to the semantic contract layer for governed work. Three posts. Five interactive simulations. Read in order, or jump to whichever part is closest to your question.
Durable execution tells you what ran. It doesn't tell you what it meant. A 3 a.m. story, an interactive demo of scattered-vs-unified evidence, and the five questions every consequential action has to answer.
Concord is not DBOS, not LangGraph, not OPA. An interactive layer stack, an audit-tier picker, and a clickable capability graph make the boundary concrete — and answer the question of what each layer actually owns.
A hotel booking traced through Concord end-to-end. Animated nine-step playback with three modes — full contract, minimal subset, no Concord at all — and a state panel that shows what each step actually writes.
Agents are starting to initiate real work. Durable execution makes that work survive failure, but durability doesn't capture what an action means, who authorized it, what it touched, what got recorded, or how to undo it. Concord is the semantic contract layer that fills that gap — a Postgres-first, runtime-agnostic library that turns commands into governed decisions, plans, effects, and audit.
Or in one line: Concord is the contract above the runtime.
The full 121-section technical spec across four tabs — Core, Multi-agent, Runtime adapter (DBOS), Domain Registry.
A longer-form positioning treatment with a nine-category comparison table.
Ten concrete pain points Concord exists to address, one per chapter.
The full design doc for the hotel-booking workflow used in Part 3 — every command, policy, artifact, and connector contract.