Product4 min read·

I Had 2 Followers and a Routine. It Turned Into Humanish.

Man typing on a laptop outdoors at night under string lights A routine, then a product.

Photo by Aathif on Pexels

Humanish started as a daily routine I ran on my own X account, on a day I had 2 followers and nobody saw my posts. The routine worked, I noticed it was repeatable, and I turned the repeatable parts into a tool. I'm one founder, it's early, and people are invited in batches. This is how it happened and the rules I decided not to bend.

Where did it start?

On July 10, 2026 my X account had 2 followers. A new account's posts reach almost nobody, and I wanted mine read. I wasn't planning to start a company. I needed a way to be seen, and I'm the kind of person who builds a thing when I need it.

What was the routine?

Four moves, every day:

  1. Reply early under people in my niche who already had an audience, with short replies worth reading.
  2. Answer anyone who answered me.
  3. Post a few originals so a new visitor had a reason to follow.
  4. Send every reply myself.

It compounded slowly, then faster. By October 4 the account was at 1,086 followers. The full week-by-week numbers, including the flat weeks, are in the numbers post.

When did it become a product?

Once the account passed 1,000 followers I noticed I was doing the same three jobs each morning: find the right posts, draft something in my own voice, and check that it didn't read like a robot before I sent it. The finding, the voice and the not-sounding-like-AI part were the hard parts, and they were repeatable. That is what a tool is for.

What shaped the quality check?

One scene. A reader answered one of my replies with "AI reply????". I had been flagged as a bot while building a tool to help people not sound like one. That is why every draft is now checked for AI tells before I see it. The full story is its own post.

What are the rules I won't bend?

  • Sound like a person, not an AI, and not like an expert. Normal people trail off and say small things.
  • A draft that fails the check is thrown out, not patched. Patched drafts keep the same shape underneath.
  • If good posts are thin, deliver fewer drafts. A short list beats a padded one.
  • Replies are never posted for you. Timing is the point of a reply, and the account is yours.
  • Real people only, not brands. Conversations, not announcements.
  • Every caught mistake becomes a permanent rule.

How does it work, simply?

You give it your X handle. It reads your public posts and builds a picture of how you write, finds fresh posts from real people, drafts replies and posts in your voice, and checks each draft. You read it, fix a word and send it. For your own posts there is an optional Buffer connection. Replies stay manual.

Is it only for my niche?

It was built for mine first, then widened. There are eight posting styles (builder, coach, analyst, commentator, creator, craft, local business and professional), each with its own idea of what a good post looks like. How the numbers behind it add up is in how to reply on X so people see it.

What went wrong along the way?

  • The voice test. I tested whether drafts really sounded like each person, using public posts from nine accounts with around 1,000 followers. The first run was no better than chance. It took four rounds of fixes.
  • A new-account test. A brand-new account showed that the connect post still used my own niche's bullets for everyone. I fixed that on October 6.
  • Payments. Indian cards can't pay USD subscriptions through Razorpay, so there are rupee plans, and the first live payment test exposed a webhook bug that is now fixed.

Where does it stand today?

The site is live and the free "Is this AI?" checker is open. You can try Humanish free for 7 days with no card, or buy Standard straight away.

What are the honest limits?

It's one account and was one niche to begin with. Funny, sarcastic and warm tones are still best effort. I can't promise growth, and I won't.

Try the checker ยท Start my free trial

FAQ

Who is behind Humanish?

I am: Snehal, one founder, posting as @boringdev77.

Does Humanish post for me?

No. You send every reply yourself, and the optional Buffer link is only for your own posts.

Is there a free trial?

Yes. Every plan starts with a 7-day trial, and the checker is free without signing up.

Will it grow my account?

I can't promise that. It's what I used on my own, and results depend on your niche and consistency.

See it work

Humanish in 20 seconds. The 94/100 and 23/100 are example scores from the same check that runs on every draft.
Read what it says

Your posts get ignored. Your replies get read. Humanish finds the right people and drafts the reply in your voice. It scans for posts worth replying to. Anything that sounds like AI gets thrown out. Adjust it in one tap. Then you press send. Two followers to a thousand, replying by hand. Humanish. Try it free for seven days.

Try it on your own account

Free for 7 days, no card needed. You send every reply yourself.

Start my free trial