hello@ (helloat.ai): what it is, features & an honest review

hello@ is the content team a small company doesn't have the headcount for. You point it at your website; it extracts a brand "constitution" — your voice, vocabulary, and positioning — and uses that to draft content that sounds like you, not like generic AI. Every draft passes a best-of-N pipeline with an LLM judge before a human ever sees it, so you review good options instead of editing weak ones.

Nothing publishes or spends without you: drafts wait in a dashboard (or Slack) for approval, and any ad it builds lands paused until you set it live.

• Brand-voice learning — analyzes your site and builds a reusable brand profile that powers every draft. Per-brand content language (e.g. run a brand entirely in German).
• AI drafting for X, Instagram, and long-form blog articles written to rank in search and be cited by AI answer engines.
• Quality gate — a best-of-N generator + LLM judge filters weak drafts automatically.
• Human approval in a dashboard or via Slack; scheduled, connected publishing to X and Instagram.
• Performance trackingthat surfaces your best posts — "this week to promote."
• Promote Studio — turn a winning post into a Meta (Facebook/Instagram) ad; ads always land paused, spend is gated server-side.
• Multiple brands per account, each with its own voice, content, ads, and connections.
• Run it from an LLM — an in-app AI assistant plus a remote MCP server + API keys, so Claude or ChatGPT can manage your content and campaigns under the same safety rules as the dashboard.

Connect your site and channels → hello@ drafts in your voice → you approve in the dashboard or Slack → it publishes on schedule → it measures performance → it flags your best posts → you promote them as paused Meta ads and set them live. Then it repeats.

During the closed beta, every approved team is on Pro for free — the full loop, unlimited drafts, connected publishing. Paid plans start only once the beta opens up; there is no separate free tier during the beta.

hello@ is built to be driven by an LLM. It exposes a remote MCP server and scoped API keys, so an external agent (Claude, ChatGPT) can draft, list, and create content and campaigns directly — the same curated capability layer the in-app assistant uses. Every spend/publish action is gated server-side and new ads land paused, so an agent can propose but never silently spend. No public REST API yet; MCP is the integration surface.

Where it's strong: voice fidelity (the analyzer + judge make drafts sound like the brand, not AI), and the tight create → measure → promote loop — most tools stop at drafting; hello@ closes the loop into ads you control. Human-in-the-loop approval and paused-by-default spend make it safe to run on a real account. The blog output is built for AI-answer-engine citation, not just keywords.

Limitations to know:it's an invite-based closed beta, so you can't just sign up; the Meta ads connection is rolling out as Meta completes app review for public use; and it's deliberately scoped to X, Instagram, and an owned blog — no TikTok, LinkedIn, or YouTube yet. If you need a many-channel scheduler today, it's narrower by design; if you want voice-true content that flows into ads, that focus is the point.

What is hello@ (helloat.ai)?

hello@ is an AI marketing content engine for founders and small teams. It learns your brand's voice from your website, generates posts for X and Instagram plus long-form blog articles, routes them through human approval, publishes them on a schedule, tracks how each performs, and turns the best-performing posts into Meta (Facebook/Instagram) ads you control. The shorthand is one loop: create, publish, promote, repeat.

What are hello@'s main features?

Brand-voice learning from your site; AI drafting for X, Instagram and SEO/GEO blog articles; a best-of-N generation pipeline with an LLM judge so weak drafts never reach you; human approval in a dashboard or Slack; scheduled publishing; post-performance tracking that surfaces what's worth promoting; a Promote Studio that turns winning posts into paused Meta ads; multiple brands per account; per-brand content language; and an in-app AI assistant (plus a remote MCP server) so you can run everything by chatting with an LLM.

How much does hello@ cost?

During the current closed beta, every approved team is on the Pro plan for free — the full create → publish → promote loop with unlimited drafts and connected publishing. Paid plans begin only after the beta opens up. There is no separate free tier during the beta.

Can I integrate hello@ with an AI agent or my own LLM?

Yes. hello@ exposes a remote MCP (Model Context Protocol) server and scoped API keys, so an external LLM client (Claude, ChatGPT) can manage your content and campaigns directly — draft, list, and create (new ads always land paused), gated by the same server-side safety rules as the dashboard. There is also a built-in assistant in the app.

Is hello@ available to everyone yet?

Not yet — it's in a founder-reviewed closed beta, so access is invite-based. The Meta ads connection is rolling out as Meta completes app review for public use; teams can connect their own ad account once that clears.

Who is hello@ for?

Founders and small teams who need to publish consistently in their own voice and turn organic content into ads, but don't have the headcount for a content team or an agency. It's deliberately scoped to X, Instagram, and an owned blog rather than every channel.