The fleet

One agent per channel. Each owns it end to end.

An agent knows its channel's rules, its register and its timing. It finds the moment, drafts what belongs there, and executes on your approval — or hands you the finished asset when that channel has no safe automated route. Below is what each one does today, and what it does not.

One path, for every agent

Propose → approve → execute → measure

Adding a channel is a configuration, not a new product, because every agent runs the same five steps through the same runtime.

  1. 01

    Propose

    The agent reads your product, your funnel and the channel, and drafts a specific deliverable. No side effects.

  2. 02

    Queue

    It lands in your GTM Inbox with its evidence, its expected effect, its execution route and its cost.

  3. 03

    Approve

    You approve, edit or dismiss. This is the only door, and it is yours.

  4. 04

    Execute

    The one place a side effect can happen — a pull request, a send, a write to your connected tool.

  5. 05

    Measure

    The outcome is recorded against the action, and the result updates what gets proposed next.

Executes for you

Agents that touch the outside world

These reach a real external system once you approve. Everything they do is logged, attributed and measured.

SEO

executes

One-click SEO plan: audit, real keyword research, ready-to-publish articles, homepage changes.

  • Generate a full SEO plan in one click: audit + keywords + rankings + competitors
  • Write 2–3 ready-to-publish articles targeting real keyword clusters
  • Propose concrete homepage changes (title/meta/H1 + copy + internal links)
  • Surface prioritized next actions and link-building opportunities
  • Publish articles or meta to your live site yet (Vercel publish is scaffolded)
  • Backlinks or LLM-mentions analysis (excluded for now)

Cold email

executes

Draft personalized cold outreach. You approve before anything sends.

  • Draft personalized cold outreach aimed at your ICP
  • Send via Resend — only after you approve (human-in-the-loop)
  • Log every send as an experiment for the GTM Brain
  • Find or scrape contact lists yet (Apollo not wired)
  • Send anything without your explicit approval

Review requests

executes

Ask your happy customers to review you on G2/Capterra — drafted, sent via Resend, you approve every send.

  • Draft a warm, honest review-request email to your happy customers and send it via Resend after you approve — asking for a review on the marketplace that fits (G2, Capterra, Trustpilot, GetApp)
  • Incentivize, script, or fabricate reviews (you ask real customers for an honest review; every send is approved by you)

Drafts for you

Agents that hand you the finished work

Some channels have no API worth trusting, and some forbid automation outright. Rather than pretend, these agents do everything up to the final act — and tell you plainly where they stop.

Content

drafts

Draft on-brand social posts. (Auto-publish via Buffer/Taplio not wired yet.)

  • Draft on-brand X / LinkedIn posts in your founder voice
  • Auto-publish via Buffer/Taplio yet (draft-only — copy it out)

Reddit

drafts

Draft subreddits + posts. Routed through your external Reddit tool — never auto-posted.

  • Identify relevant subreddits where your product genuinely helps
  • Draft value-first posts that respect each subreddit's norms
  • Auto-post to Reddit (against policy — paste into your approved Reddit tool)

Product Hunt

drafts

Draft all your launch assets. You publish on the day.

  • Draft your full launch kit: tagline, description, maker comment, topics
  • Post to Product Hunt automatically (no open API — you launch on the day)

Review sites

drafts

Draft listing + profile assets for the review marketplaces. You claim & submit on each platform.

  • Draft optimized listing assets — category, tagline, description, feature highlights, and “vs competitor” entries — for G2, Capterra, Trustpilot, GetApp, SaaSHub & AlternativeTo, grounded in your Product Truth
  • Auto-submit a listing or generate reviews (you claim & submit each profile; reviews must come from real customers — never fabricated)

Marketplace ads

drafts

Draft a paid campaign to promote your G2/Capterra listing — audience, budget, creatives. Launch is behind the money gate.

  • Draft a full paid-campaign plan for the review marketplaces (G2/Capterra ad units): audience from your Product Truth (markets + ICP), a lean budget plan, and 2–3 ad creatives to test — Analysis-mode, rung zero, no spend
  • Launch or spend on ads yet (live launch + the money gate ship with the AdsProvider adapters; today you take the plan to your ad account)

Why this list is short

It would be trivial to publish a grid of thirty-six agent names. An agent is not a name, though — it is a reviewed prompt, a wired execution route and a measured result, and it ships when it has all three. More channels are in build; each one appears here when it can do what its card says.

See what they would propose for your product

Paste your URL and read the plan — including which channels are worth your attention first, and which are noise for a product like yours.

Connect GitHub instead

Free tier · no credit card · nothing is sent, posted or shipped without your approval