Everything people ask before they connect WhatsApp.

No. auto-reply.to is web-only — sign in, scan the QR (just like WhatsApp Web), and you’re done. Nothing runs on your phone, nothing in your browser, nothing to keep open. Your assistant lives in the cloud and replies even when your devices are off.

No. auto-reply.to works with your regular personal WhatsApp by linking the same way WhatsApp Web does — scan a QR and you’re in. No Business account, no Meta Cloud API, no Twilio, no developer setup of any kind. If you already have a Business account that’s fine too; auto-reply.to simply doesn’t require one.

Solopreneurs, freelancers, indie founders, coaches, course creators, and anyone running a one-person business who’s drowning in client DMs. If you’re a team, auto-reply.to still works — but the product is designed first for the person doing everything themselves.

Every plan starts with a 7-day free trial. You pick a plan up front, but you’re not charged until the trial ends, and you can cancel any time before then from the billing page.

auto-reply.to reads and replies on a low-frequency schedule (every 60s) with a humanistic delay before each reply, so usage looks similar to a careful human. That said, WhatsApp’s ToS does not formally permit any kind of automation on personal accounts, so we recommend linking a secondary number for maximum safety. auto-reply.to is at-your-own-risk.

No. auto-reply.to only fetches messages from chats you explicitly toggle on in the dashboard. Everything else in WhatsApp is invisible to us.

Yes. Every reply ends up in your audit log with the hidden reasoning chain that produced it. You can also override or pause anything you don’t like.

No. Transcripts are fetched to score a contact, then dropped. We keep only derived data: scores, short reasoning, and engagement signals — the same privacy posture as auto-reply.

From the persona prompt and the encrypted knowledge base you fill in — those are static facts. On Enterprise, you can also connect HTTP tools so the assistant fetches live data from your APIs during a reply instead of guessing. If a question isn’t covered by either, it sends a natural deferral like ‘let me check and get back to you’.

On Pro and Enterprise you can turn on “Read media” per chat. Images and PDFs are sent to the vision model for that reply only — we never store the bytes. Voice notes and videos are not transcribed in v1; the AI only sees a short placeholder like “user sent a voice note”.

We use a frontier reasoning model from a top-tier AI lab — the same class of model powering the smartest products you use today. Pro users can opt into our premium reasoning tier on a per-chat basis when persona fidelity matters most. We swap to the best available model as the field evolves, so you always get the strongest engine without having to think about it.

Campaigns scan WhatsApp chats you already have and grade each contact hot, warm, or cold against an offer you describe. We draft openers in your voice — we never invent leads or scrape strangers.

We tell you straight. If a scan finds only cold contacts, you’ll see an honest empty state — no fake urgency, no invented pipeline.

One outbound message sent by your assistant counts as one reply. Draft openers you never send don’t count, and neither does scanning chats for a campaign — you’re only charged credits when a message actually goes out.

Yes. Upgrade, downgrade, or cancel any time from the billing page in your dashboard. Changes take effect at the end of the current billing period, and there’s no cancellation fee.