Posting Mental Health Content Creates an Inbox You Owe

This is for an account that publishes rather than practises: an awareness page, a quotes account, a writer, a student-run collective, the content arm of a company that is not in healthcare. Somebody who makes carousels about burnout, or reels about intrusive thoughts, or a weekly post that a lot of people save.

If you are a clinician, the therapists post is the one you want. If you coach non-clinically, that has its own build. This page is about the account with no practitioner behind it at all, which is the commonest case and the least written about.

Good content on this subject produces a disproportionate inbox

A post about a leaking tap gets comments about taps. A post that accurately names how somebody has been feeling for eight months gets a different thing entirely, and it does not arrive in the comments.

It arrives in DMs, at night, from people who have decided — reasonably, on the evidence available to them — that whoever wrote that carousel understands them and might be able to help. Some of those messages are heavy. A few are urgent. None of them are about the carousel.

That is not an edge case in this niche. It is the normal consequence of the content working.

And the account is almost never staffed for it

Here is the honest bit. On most pages like this, the entire operation is one person with a laptop, no training in this area, no rota, no supervision, and a full-time job elsewhere. There is no on-call. There is nobody to escalate to.

The rule everywhere else in this series holds here and is absolute: when a difficult message arrives, the automation sends nothing and a human sees it immediately. But this account has to answer the question that rule assumes — which human, and when are they awake — before building anything. If the answer is "me, sometimes, when I remember", the design has to account for that rather than pretend otherwise.

Do not invite an inbox you cannot answer

This is the lever that is actually available to a publisher account, and it is a content decision rather than a configuration one.

An automation only fires because somebody typed a word, and that word only exists because you asked for it in a caption. So you control which posts open a channel.

Say what the account is, where people read it before they write

An account like this has no name, no face and no stated role, which means readers fill that in themselves and they fill it in generously. The correction has to be published, not sent.

In the bio and in a pinned post, in plain words: this account shares writing and resources, it is not a service, messages are not monitored continuously, and it is not a route for urgent help. Then, in the same place, the route that is — services relevant to where your readers actually are, chosen by you, checked twice a year. This post will not name any: they differ by country and they go stale, and a wrong number is worse than no number.

None of that belongs in an automated reply, where it arrives after somebody has written and reads as a door closing.

The post that travels brings a different audience

When something breaks out of your follower base, the composition of who is writing changes before the follower count does. More first-time senders. More people with no context for the account. Sarcasm, and hostility, and messages whose tone you cannot read.

Three settings matter more on that day than any other. First-time senders only, narrowing the automation to genuinely new people. Quiet hours, so nothing goes out overnight into a spike you are asleep for. And the negative keyword list, which on a post that travels is doing more work than the trigger is.

Every reply also passes a fixed set of checks before it sends — takeover, window, dedupe, cooldown, quiet hours, rate budget, credits, content safety among them — and a blocked reply is recorded with its reason in Activity. Read that screen the morning after a post travels.

The one pile that is entirely safe

There is a category of message this account gets that nothing else does, and it is genuinely templatable: permission and credit. Can I repost this. Can I translate it. Can I put it in my school presentation. Please credit the original artist. Someone has copied your post.

Write those answers once. They are administrative, identical for everybody, and they may be a large share of your volume. The same goes for "where is the newsletter", "is there a print version", "which font is that".

That is the honest automatable set on a mental health publishing account: logistics and rights. Everything about a person stays with a person.

The costs, and what does not exist

Templated replies are free and unlimited, so the set above runs at no cost. A credit is spent only when the AI writes a new reply. Free tier is 100 credits with no card, packs from ₹499.

We do not make content: no captions, no scheduling, no publishing, no hashtags, no analytics, no follower-growth feature. No WhatsApp, no broadcasts, no sequences. Leads leave as a CSV. Official Graph API only, and nobody can message a person who has not messaged first.

Takeover means the automation stands down in a thread permanently the moment you reply by hand. On this account, expect to reply by hand to almost everything that matters. If the workload of that is the actual problem, the burnout post is more use than any configuration on this page.