Skip to main content

AgentForge

AgentForge is the landing page and cabinet prototype for Hermes Agent. It is the commercial shell around a Hermes deployment: the public face, the onboarding path, the login layer, and the place where the product becomes something a client can actually buy and use.

What AgentForge is for​

  • turn a Hermes deployment into a product
  • give the agent stack a clear public entry point
  • keep onboarding, access and lead capture in one flow
  • separate the commercial layer from the runtime layer
  • make deployment repeatable instead of custom every time

What is inside​

Under the hood, AgentForge is built as a practical web stack:

  • a React + Vite client
  • an Express backend
  • PostgreSQL + Prisma for data
  • Google and Telegram auth paths
  • Turnstile for bot protection
  • Docker / Ansible deployment automation
  • SeaTable sync for leads and source-of-truth handling

How to use it correctly​

  1. Start with one clear AI-agent use case.
  2. Deploy the Hermes runtime in a dedicated stack.
  3. Use AgentForge as the customer-facing shell, not as a generic website builder.
  4. Keep the first flow simple: entry, auth, demo, contact, deployment.
  5. Save leads and status in SeaTable so the data does not split across tools.

What not to do​

  • do not confuse AgentForge with Hermes itself
  • do not treat it as a general CMS or page builder
  • do not keep the source of truth in spreadsheets
  • do not add extra channels before the main onboarding flow works
  • do not bury the product value under infrastructure details

In one line​

AgentForge turns a Hermes deployment into a product people can understand, try, and pay for.

Runtime layer

For the agent runtime itself, see Hermes.