How to log Huel and Orgain by voice

· voice-tracking, macros, protein, ai

How to log Huel and Orgain by voice

How to log Huel and Orgain by voice

You’re trying to track macros. You tap the mic on your tracker, you say “two scoops of Huel, one orange,” and the transcript comes back: “two scoops of fuel, one orange.”

You delete “fuel,” type “Huel,” hit estimate. Tomorrow morning, same thing. Same scoop, same orange, same wrong word.

This is the universal voice-tracking experience. Apple’s speech recognizer is excellent at common English, mediocre at fitness lingo, and absolute garbage at brand names. Orgain becomes Oregon. Barebell becomes barbell. Mezcla becomes mescaline. Anything niche, anything new, anything outside the training corpus.

Most apps just shrug. Some bury a “report this mishear” button in settings that asks you to do the speech-tech team’s job for them. TrakMac took a different angle. We just read your edits.

Why voice macro trackers mishear protein brands

Open a regular macro tracker. The voice feature is usually a thin wrapper around your phone’s keyboard dictation. You speak, the transcript appears, the macro estimator runs on that text. End of pipeline.

If the recognizer botched your input, the estimator botches the output. Garbage in, garbage out. So you either retype the brand every time, or you give up and switch to manual entry.

The deeper problem: the recognizer doesn’t know what you eat. It knows English. You eat brand-name protein.

That gap matters because protein supplementation isn’t optional for most people training seriously. The International Society of Sports Nutrition recommends 1.4 to 2.0 grams of protein per kilogram of body weight per day for active adults, with higher intakes when the goal is preserving lean mass during a calorie deficit (ISSN position stand). For most people hitting those numbers, branded powder or bars are how you actually get there. A voice tracker that can’t recognize them is a voice tracker that doesn’t work.

How TrakMac handles brand names

Two layers.

The first is a curated vocabulary. TrakMac ships with brand names that fitness folks actually use. Huel, Orgain, Mezcla, Barebell, Dave’s Killer Bread. These get fed into Apple’s recognizer as contextual strings before you hit record, so the engine has a fighting chance the first time you say “Huel.”

That covers the brands we already know about. The second layer covers the ones we don’t.

When you say “two scoops of fuel” and edit it to “two scoops of Huel” before tapping Estimate, TrakMac quietly captures that fix. Two corrections of the same mishear is enough. After that, the recognizer is biased toward “Huel” specifically for your voice in this app. You won’t have to edit it again.

No “tap to report” button. No settings panel. No telling the team about your transcript. Just edit and move on.

What about brand-new words?

You opened a tub of some powder you found at the gym front desk that nobody has heard of. You log it once, fix the mishear, log it the next day, fix it again. Third log: it lands clean.

The system doesn’t need the brand to exist anywhere else in the world. It only needs you to use it twice and prefer your spelling. After that, it’s part of how the app understands you. The next stranger word you start eating gets the same treatment.

This matters more than it sounds. The supplement aisle gets new entrants every quarter. Last year’s go-to powder is this year’s discontinued SKU. A voice tracker that depends on a pre-baked dictionary of brands has a permanent staleness problem. One that learns from you doesn’t.

When this matters and when it doesn’t

It matters whenever your routine involves products the recognizer doesn’t know. Most lifters fit. Most cyclists fit. Anyone on the keto, GLP-1, or Whole30 path stockpiling niche brands fits. Anyone who eats the same five things on repeat fits double, because two edits is the whole training set.

It matters less if your meals are mostly whole foods. Apple already handles “two eggs and a piece of toast” without help, and the learning loop quietly sits idle.

It mattered for me because I drink Huel five days a week. I got tired of saying “fuel” and being misunderstood, on my own phone, by my own app. So we fixed it.

Download TrakMac

TrakMac is live on the App Store. Voice-first macro tracking that learns the brands you actually consume, not the ones the speech recognizer was trained on a decade ago. Free 7-day trial, no card required. Check the FAQ for what to expect.

Download TrakMac free.

Download TrakMac free →