Agents That Pay for Themselves

Five agent builds mapped to what businesses actually pay for — each one stops where your judgment starts.

Tested builds for Hermes Agent, verified on v0.17.0/v0.18.2, July 2026. Nothing sends, publishes, or imports without a human's approval.

Someone scraped 1,000+ AI automation jobs posted on Upwork over six months and sorted what businesses put real budgets behind. No robots-run-everything fantasies made that list. Five boring jobs did, over and over: lead generation, outreach qualification, content on a schedule, review handling, inbox response.

One buyer in that dataset said it in a sentence: "I'm drowning in lead gen, I need this to run on autopilot."

This kit is five Hermes Agent builds aimed at exactly those jobs — the configs, prompts, cron specs, and operating procedures, tested and documented for a non-technical business owner who can paste a command and read the result.

THE FIVE BUILDS:

Lead Scout Pro — researches and enriches a draft lead list overnight. You review over coffee; nothing gets contacted or imported until you say so.

Outreach Drafter — qualifies your approved leads and drafts personalized cold outreach. It never sends. You do, from your own account, in your own name.

Content Engine — drafts a full SEO article on a cron, in your voice. Abstracted from a content system we run in production for our own properties. A human edits, a human publishes, every time.

Review Responder — a same-day, personal draft reply for every review, hard ones flagged first. You post them.

Inbox Lead Responder — turns a missed 11pm inquiry into a staged draft reply within about two minutes, ready to approve from your phone. One optional automode exists and it can only send a neutral "got your message" acknowledgment — it ships OFF.

That last pattern is the whole product: every build does its job up to a hard line — anything that touches money or gets sent waits for a human. That boundary is why these run for months instead of blowing up in week one, and it's the first thing to show anyone who's nervous about "AI running the business."

The hub guide also includes a payback worksheet you fill in with YOUR numbers — what a lead, a review, or an article is worth in your business versus what the agent costs to run. No invented dollar figures anywhere in this product. If a build doesn't pay for itself in your own numbers, the guide tells you not to run it.

And if you'd rather install these for OTHER businesses: the included service appendix covers packaging, setup-fee-plus-retainer pricing frameworks, client scoping questions, and the delivery checklist — built honest, no income promises for you to make or receive.

WHAT'S INCLUDED:

Designed PDF hub guide — the five builds, the payback worksheet, the boundary chapter, honest running costs

5 build folders, 5 plain-text files each: README, agent config, cron spec, approval boundary, test checklist — copy-paste ready, straight quotes throughout

SELL-IT-AS-A-SERVICE.md — the pricing appendix for selling these as services

CHANGELOG.md — dated, with the exact Hermes versions tested against

Living Notion version, updated as Hermes updates

Lifetime updates — buy once, re-download forever; build #6 will be added to this same product

HONEST SCOPE: $67 is the launch price. It rises when build #6 ships — that's the only reason, and existing buyers get build #6 free as an update. This kit assumes you already have Hermes Agent installed and responding (our free guide gets you there if not). It is not an installation tutorial and not a course. It contains no income claims: what these builds earn depends on your business and your follow-through, and the guide hands you a worksheet, not a projection.

Unofficial. Not affiliated with Nous Research.