how it works · v1.0

launch in three steps. then we keep working.

the launch is the on-ramp. discoverability is the job — and the job is never done.

  1. 01
    connect

    paste a repo. or a url. or skip both.

    We don't need code. We need to know what your agent does and who it's for. If you have a GitHub repo, paste it and we'll save you typing. If your agent lives on Replit, Bolt.new, Lovable, v0, or a custom GPT link — paste that. If neither — we'll ask.

    • github· public repo, read-only
    • url· any hosted agent
    • none· interview only · 8 minutes
  2. 02
    onboard

    a short interview. plain english.

    Three minutes for the basics — agent name, what it does in one sentence, target audience. Then a magic moment: your first brand kit preview. Five more minutes for the rest — voice, FAQs, distribution targets. We use your answers to write your launch copy and pick a brand kit.

    • 5quick questions · 3 min
    • 7–10deeper questions · 5 min
    • auto-saves· come back anytime
  3. 03
    launch

    your URL, live.

    Your subdomain goes live with SSL. While you do nothing, we ship the AI-readable layer underneath — and start submitting you across eight marketplaces.

    • domain· [your-agent].envvoy.ai · SSL ready
    • ai layer· llms.txt · schema.org · sitemap · structured Q&A
    • submissions· 8 v1 marketplaces, ongoing
    github topicssmitheryagent.directoryvirtualsa16z listawesome-ai-agentsproduct huntai agent marketplace
and then we keep working

launch is fast.
discovery is forever.

every day, we check the AI engines for citations — chatgpt, claude, perplexity, and more. once a week, on monday morning, you get a summary email: what changed, who cited you, one suggested action. the work compounds.

monday · 7:02 am
illustrative exampleweek 4 · kestrel

your visibility went up +12 points this week.

visibility score68+12
new citations+21across 3 engines
recommendation rate42%+8
one thing to do this week

add three usage examples to your FAQ — perplexity is missing context for refactoring queries.

what we learn

the more agents we manage, the smarter we get.

because we run discovery for many agents, we see patterns no individual builder ever could — which directories are actually converting this month, which copy formats AI engines are citing more frequently, and how ranking behavior shifts as chatgpt, claude, and perplexity update their models.

that intelligence flows back into every agent we manage. through the monday briefing. through the AEO wizard recommendations. through continuous updates to how we submit and position your agent as the landscape changes.

you get the benefit of what works across the whole ecosystem — not just what you could figure out alone.

most asked

three questions builders ask first.

do I need to know how to code?

no. if you can describe what your agent does and who it's for, we can launch you. no terminal, no github, no code — you bring the idea, we do the launch.

what if my agent isn't on github?

paste any url where it lives — replit, bolt.new, lovable, v0, a custom gpt link. or skip the url entirely and we'll do an interview-only onboarding.

what happens to my site if I cancel?

you keep your domain. you can cname it elsewhere or take the html with you. we don't hold anything hostage.

see the full faq

every agent deserves a chance to be heard.

we work for you every day.

no card neededfree during beta