Set Up Your Supplement Stack in 30 Seconds with AI

The fastest way to abandon a health app is a 20-minute onboarding flow. Most supplement trackers make you add each supplement one at a time, searching a database, selecting the form, entering the dose, choosing a time block. Do that for 10 supplements and you've already lost patience.

Stack Almanac takes a different approach: just tell it what you take.

How it works

When you sign up, you'll see a text box. Type or paste your supplements however you naturally describe them:

Morning: vitamin D 4000IU, fish oil 2g, B complex Before bed: magnesium glycinate 400mg, zinc 25mg, L-theanine 200mg

Or even less structured:

I take tyrosine, NAC, berberine in the morning on an empty stomach, then D3 and K2 with breakfast, magnesium before bed

The AI reads your text and parses it into structured supplement entries, matching each one against the supplement catalogue with the correct name, form, dose, and time block.

What happens next

After parsing, you'll see your supplements organised into three categories:

Matched (green): the AI confidently identified the supplement, form, and dose. These are ready to add to your stack with one tap.

Ambiguous (amber): the AI found a match but isn't sure about the form or dose. For example, "magnesium" could be glycinate, citrate, threonate, or oxide. You'll see options to clarify.

Unmatched: the AI couldn't find a match in the catalogue. You can request it be added, or specify it manually.

Most stacks parse in under 5 seconds. A typical 10-supplement protocol goes from plain text to a fully structured, timed, interaction-checked stack in about 30 seconds.

Instant interaction checking

The moment your stack is created, Stack Almanac automatically checks for interactions between your supplements. Before you've logged a single day, you'll already know:

• Which supplements conflict or compete for absorption
• Which ones need to be taken at different times
• Which combinations are synergistic
• The severity and evidence grade for each interaction

This is day-one value, useful information delivered before the app asks anything from you.

Why this matters

The traditional supplement tracking onboarding goes like this: search for a supplement, select from a list, enter dose, choose form, pick a time, save, repeat. For someone taking 12 supplements, that's 12 repetitions of a 30-second process. Six minutes of tedious data entry before they see any value.

Stack Almanac's natural language setup inverts this: value first, then refinement. Get your stack in, see your interactions, and start benefiting immediately. Refine the details later.

What it's powered by

The parsing runs on Claude's AI, called server-side from a Supabase Edge Function. Your plain text is matched against a curated supplement catalogue with over 200 entries covering vitamins, minerals, amino acids, adaptogens, nootropics, herbs, and more.

The AI considers context clues: "before bed" maps to the Wind-Down or Sleep time block, "with breakfast" maps to Breakfast, "fasted" maps to the Fasted block. Doses are normalised and validated against typical ranges.

If the AI can't parse something with confidence, it asks rather than guesses. Accuracy matters more than speed when it comes to your health data.