← All work

Case study

Tackie

Most goal apps fail the people who need them most: the ones frozen by a big goal, or who fall off after one bad day. Tackie asks one question at a time until "get fit" becomes something you can start this week, plans the week to fit your time, and never breaks your streak for missing a day.

Product
Pinn.Media, in-house
Year
2026
Stage
Beta, free to start
Stack
Next.js, Supabase, Claude
Live
tackie.io
  • Product
  • Brand
  • UI/UX
  • AI engineering
  • Illustration & animation
  • Web app
  • SEO
Visit tackie.io ↗

Tackie is Pinn.Media's own product, an AI goal planner. You describe a goal in plain words; it sharpens the goal through a short conversation, breaks it into a weekly plan scaled to your pace, and shows your progress as a mountain climb that adapts every week to how the last one went.

Why we built it

We wanted proof of what we sell, end to end, with nobody else's brief in the way: a brand, an app, an AI pipeline that is safe and cheap enough to offer free, and a site that brings people to it. Tackie is that proof.

What we built

Started as a proof of concept called Ascent in mid-June 2026 and renamed Tackie the next day, the core product came together in the following fortnight across nearly four hundred commits:

  • the one-question-at-a-time sharpening conversation with safety screening
  • pace-scaled weekly plans placed in your free time, and the weekly reckon that rewrites the next week
  • the climb, an Ascent Score, forgiving streaks, a Pomodoro timer and a "lighten this week" rescue
  • accounts with Google sign-in and two-factor authentication, row-level security on every table
  • calendar links, an .ics feed, web push reminders, a no-account try-it mode
  • the landing site, blog, changelog and admin tools

Outcome

Tackie is live in beta and free to start. Native iOS and Android apps are on the way; today it installs as a web app. We are not quoting user numbers: it is early, and the landing page says so ("join the first climbers").

AI

Tackie, the planning engine

A conversation that asks exactly one question at a time, a weekly plan that scales to your pace, and a Sunday check-in that rewrites next week from what actually happened.

Sharpen. You type a goal however it comes out. Tackie asks one question at a time, the model decides how many, until the goal is specific enough to start. A safety screen runs first: an unsafe target (say, a crash diet) is gently reframed toward a sustainable pace, and an impossible one is talked down to a real first step.

Plan. Each week is three to five tasks, each one a single sitting you can tick yes or no, starting with whatever is in your way. How many and how hard depends on the time you have and the intensity you chose, and they are placed in your free windows.

Reckon. At the end of the week you say how it went, Proud, Steady, Tired or Rough, and next week is written from that, not from the plan you had. A missed week makes the trail longer; the mountain only moves forward.

Under the hood the heavy reasoning uses a larger Claude model and the quick jobs (planning, safety, insights) a smaller one, so the product stays affordable to run for free. Statistics are computed in code and the model only narrates them, so the numbers are always right.

Try page: What's a goal you've been putting off?
"What's a goal you've been putting off?" No account needed to start; the draft stays on your device.
See it in motion: watch a vague goal become a plan
"Watch a vague goal become a plan." The landing page's live preview of Sharpen, Today, Climb and Calendar.

Animation & 3D

Tackie, the climb

Progress as a painted mountain: day and night with your local clock, clouds that clear as the week gets done, and a climber who reaches the summit when you do.

A progress bar tells you what you have not done. The climb shows you what you have. The scene is built from painted layers, sky, mountain, ground, sun or moon, clouds and a climber, animated with plain CSS so it runs smoothly on any phone. It turns from day to night with your clock. Finishing this week's tasks brightens the sun and clears the clouds; overall progress walks the climber along a trail traced from the painting; the summit gets a victory pose and a burst of colour. Under reduced-motion settings it all holds still.

Climb scene in daylight with the climber on the trail
Day: the climber on the trail, clouds still in the sky.
Climb scene at night under the moon
Night, by your local clock. The mountain only moves forward.
Climber celebrating at the summit
The summit. The goal you described in week one, done.

Web

Tackie, the landing site

A playful, honest landing page: who it's for, a live preview, the plans, and a blog on actually reaching goals.

The landing page names the four people Tackie is for, those frozen by big goals, those who start strong and fizzle, people with ADHD who need the structure held for them, and anyone who quits after one missed day, and says plainly that it is not medical advice. A FAQ covers price, data privacy and how it differs from asking a chatbot. The blog publishes practical articles on momentum and setbacks, and the site describes itself to AI assistants in llms.txt.

The brand is Gen Z and warm: electric violet with mint, banana and pink used sparingly, Bricolage Grotesque headings, and a rule that no screen ever uses every accent at once.

Tackie landing page: Turn I want to into actually doing it
"Turn 'I want to' into actually doing it." Type a goal straight into the hero.
Who it's for: You're not bad at goals. The tools were.
"You're not bad at goals. The tools were." Four people the product is built for.
FAQ: Everything you might be wondering
Price, privacy, and how this differs from just asking ChatGPT or Claude.
Tackie blog: Goals, plans, and the climb
The blog: honest writing on momentum, setbacks and breaking big things into small ones.
Tackie landing page on a phone
On a phone, where it is meant to live. Native apps are on the way.

One team, from identity to intelligence.

Brand, software, AI, and growth from a single team, not separate vendors. That is why the positioning, the copy, the site and the data model on this project agree with each other.

Book a discovery call