Press

Everything here has been checked.

If you are writing about us, take the numbers from this page rather than from a deck or a summary. These are the figures we can substantiate, and we will tell you when one is out of date.

Press contact hello@tastetwins.app · put “Press” in the subject and we will move you up the queue

The facts

  • CompanyLOVE SALONE, L.L.C.
  • ProductTasteTwins — native iPhone app
  • FounderCorey Arnold, founder and chief architect
  • BasedWashington, D.C. area
  • StatusPrivate testing with invited users. Not on the App Store.
  • Canonical places1,475 (30 August 2026)
  • Menu items described956 (30 August 2026)
  • Photographs2,005 (30 August 2026)
  • Websitetastetwins.app
  • Contacthello@tastetwins.app

Boilerplate

Use either. Both are cleared.

One sentence

LOVE SALONE, L.L.C. builds TasteTwins, an iPhone app that learns what a person likes at the level of the individual dish rather than the cuisine, and uses that to connect people whose taste genuinely overlaps.

One paragraph

LOVE SALONE, L.L.C. is an applied AI company in the Washington, D.C. area. Its product, TasteTwins, models food at the level of the dish — describing each one by its sensory properties rather than filing it under a cuisine — and learns a person’s palate from how they rate what they eat. The approach was shaped by the founder’s Sierra Leonean background and by the observation that food recommendation systems trained on Western restaurant data lack the vocabulary to describe much of the world’s cooking accurately. The company is developing its taste model against a seed dataset of West African dishes, on the principle that a model rich enough to describe cassava leaf stew is more than rich enough for everything easier. The app is in private testing.

Assets

  • TasteTwins wordmark (PNG, transparent). It is dark; on a dark background, invert it — that is what this site does.
  • App icon (PNG).
  • Interface screenshots: the images on the home page and its colour gallery are genuine captures from the current build on an iPhone 16 Pro simulator, taken 31 August 2026, showing sample content. Higher-resolution originals on request.

You may use these to illustrate coverage of the company. Please don’t alter the wordmark’s proportions or colours beyond inverting it for contrast, and please don’t use our marks in a way that implies we endorse a product or publication.

Claims we ask you not to make

We keep an internal register of what we can and cannot substantiate, and the following have all been asserted about products like ours before. None of them is true of this one.

  • It is not genetic. “Taste DNA” is our name for learned preferences. There is no DNA test, no sample and no genetic data. This is the single most common misreading and we would be grateful if you did not repeat it.
  • It is not a medical or clinical tool and does not diagnose, treat, or prescribe a diet. It has not been through a designed clinical study and we do not describe it in health-outcome terms.
  • It is not validated. Matching works and testers find it useful. We have no measured accuracy figure to give you, and any number presented as one did not come from us.
  • The catalogue is not in the millions. The real figures are above and they are modest. Earlier company material overstated this; the numbers on this page are the corrected ones.
  • Nothing is anonymised beyond what we claim. Taste data is protected by access controls and encryption, not by cryptographic anonymisation. The privacy policy is precise about this and it is worth reading before writing about our privacy posture.
  • No research data has been shared. The academic and public-health programme is a design goal. No dataset has gone to any partner.

The demo

The interactive demo on this site is a browser simulation with invented people and venues and deliberately simplified scoring. It is a good way to see the interface and a bad basis for describing the product’s capability. If you want to see the real thing, ask and we will walk you through it.

Interviews

Corey Arnold is available to talk about dish-level taste modelling, why food recommendation systems fail on non-Western cuisine, and building consumer AI with a defence-sector security background. Email hello@tastetwins.app.

Found something on this site that contradicts something else we have published? Tell us — we would rather correct it than have you work around it.