I think it's a pretty ambitious vision for a spectrometer. The one they show is acknowledged to be a "mockup". My guess is that they plan to identify what kind of food it is from the spectrum, then look up the nutritional facts from an existing database they create. That's already very ambitious -- if they had proposed actually scanning and identifying proportions of, say, potassium, that would be even harder, maybe impossible.

It is supposed to be a Raman spectrometer. What are the capabilities of that technology with 1 stimulus frequency? This is beyond my knowledge.

Did it say that in the video or on the page? It's hard to evaluate closed-source claims, and their video is very markety/dumbed down.

The text says Raman spectrometer.

They've posted a 2013 product update, and it looks promising, maybe too promising?

I looked over some papers regarding gluten detection and saw that the main interest in that field has more to do with Melamine detection in gluten, be it in animal feed or baby food, it has a noticeable spectral footprint, some links on this subject:

2013 update: http://tellspec.com/howitworks/

Applications of Vibrational Spectroscopy in Food Science http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0470742992.html also in googlebooks: http://books.google.co.il/books?id=MTLoO-IJJYIC&pg=PA655&lpg=PA655&dq=SERS+Gluten&source=bl&ots=dqHnwS5LZl&sig=KHR_r0u0WIhjH7uGmvG9zvamh6E&hl=en&sa=X&ei=mt77UozoCMvRsgb6jICYDQ&redir_esc=y#v=onepage&q=Gluten&f=false

paper on melamine near Gluten: http://afrsweb.usda.gov/SP2UserFiles/person/3003/2009LiuEtAl63-4ApplSpectroscopy477-480.pdf