HHermesAIDocs

Documentation overview

Where to start, what is covered, and how the HermesAI docs are organized.

HermesAI is an AI newsroom built for modern media. It monitors a 1.400+ source catalog continuously, clusters the same story across outlets, cross-checks every claim, drafts in your language, and routes drafts through editorial review before delivery into your CMS.

This documentation site is organized around the product that exists today — not a hypothetical future stack.

Start here

What these docs cover

Development

How to boot the project locally, which services it depends on, and which commands are considered canonical.

Product

How the tenant newsroom and platform admin experiences are split, which workflows belong in each domain, and how HermesAI treats the localized article as the primary object.

Architecture

How the current Next.js 16 application uses Drizzle-backed Postgres, Redis-backed runtime state, the source catalog pipeline, and AI runtime policy controls.

Operations

How to reason about webhooks, catalog syncing, and other operational workflows that sit adjacent to the product UI.

Integrations and commercial posture

What the WordPress connector does today, how pricing is framed under the v2 cost-envelope model, and which public-site pages are currently implemented.

Core product model

DomainAudienceRoutesResponsibilities
Tenant newsroomNewsroom users/app/*Feed, story workspace, compose, refinement, editorial review, monitoring, workspace settings, notifications
Platform adminPlatform operators/admin/*AI runtime config, diagnostics, review operations, sources, queues, costs, deliveries, tenant oversight

Documentation discipline

HermesAI already has enough depth that documentation drift becomes expensive quickly.

When the route model, billing rules, provider support, webhook behavior, or public-site posture changes, update the affected documentation in the same change set.

On this page