The real question is where do you plan to sell this?
That is the first determinant of what you need to worry about. Additionally, you do not
technically need to comply with any standard set by UL/TUV/CSA/ETL/MET or any other non-government entity. Their "standards" do
not represent federal laws, specifically for the U.S. that would actually encompass the statues/acts enacted by Congress overseen by the commissions granted authority to enforce them. The relevant commissions for the U.S. in terms of a wi-fi device would be the FCC, CSPC and potential overlap (for installation and usage methods) into OSHA and
maybe even the EPA (depending on the device). As you can see plainly here, as far as product safety goes there are
zero explicit regulations on (for example) power supplies from the CSPC:
http://www.cpsc.gov/en/Regulations-Laws--Standards/Regulations-Mandatory-Standards-Bans/Now,
this says nothing of state and local governments, which are also free to regulate any litany of wacky/random/variegated things at all, except where possibly prohibited by federal laws which would supersede or long-respected precedent of some kind. Most people assume that because you see UL/TUV/ETL listing everywhere (and normally it's a composite Canada and US listing) that they are somehow "required". It's simply not true. The reason they are so ubiquitous is that major distributors, resellers, retailers and manufacturers have
very strict rules about what kind of certification standards they will require for product safety. At the end of the day, Walmart and the person/manufacturer/hobbyist selling the dangerous device are the people bound to get sued so, yes... for explicitly or even just potentially dangerous stuff it's well within their best interest (as a large retailer or manufacturer) to have a boatload of relatively strict certifications by respected 3rd party testing companies slathered all over the product label.
Dealing with the FCC is a bit different. The only way you can potentially avoid an expensive test is if your equipment for manufacture falls under needing only "verification", which means as long as you can effectively prove that the equipment meets the requirements for the particular subsection or specific electronics type it falls under, you're good. This can be done through any non-accredited lab (this is relatively important because, well... how do you define a "lab"? My R&D room probably qualifies as long as my certification schedule looks dandy I would bet...). Equipment requiring a DoC (Declaration of Conformity) gets a bit trickier because now you have to use a lab that's accredited by the NVLAP, A2LA or MRA. Luckily, for both these situations nothing needs to be submitted directly to the FCC to keep on file. It's squarely on you (good luck!)
The tough part is when you have to actually certify your product with the FCC. That requires explicit filing with the FCC or a Telecommunications Certification Body (TCB). There's a ton of really good information (surprise) directly from SparkFun on this stuff:
https://www.sparkfun.com/tutorials/398Hopefully, in this situation your device simply falls under subpart
B of part 15 of title 47 of the Code of Federal Regulations, which is an unintentional radiator. Being that the OP is describing a radio/wi-fi/whatever, this is sadly not the case. That's governed by subpart
A, which (lucky you) is the most restrictive regulation you can fall under for a consumer electronics product (which makes perfect sense because the FCC's whole job is watching over radio-spectrum bandwidth usage). You can typically get away with a cert cost of around $1k for an unintentional radiator, but an intentional one is more in the realm of $5k+.
Now CE is something else altogether. People who are talking about them in the same sentence as if they were somehow in
any way homogeneous have not seen this process happen. CE has all sorts of different directives. Here's a tidy little list:
http://www.newapproach.org/Directives/DirectiveList.aspIf you ship a product with some
other type of already certified power supply, you can normally get away with avoiding the low-voltage directive (2006/95/EC) which makes things quite a bit easier. Now you
absolutely no matter what have to still worry about RoHS (directive 2011/65/EU), possibly RTTE (directive 1999/5/EC) and definitely EMC (directive 2004/108/EC). My knowledge on CE is not quite as good but I believe it is common to be able to put the CE label on a device if you are quite sure that it abides by the pertinent directives and while explicit certification by an accredited body is "awesome", it is not typically critical. I could be wrong on this. All the products I oversee have explicit CE
certification for various directives (also spendy) so I've not dealt with trying to subvert utilizing a lab here.
But wait! Are you going to ship to Mexico or Japan or Australia or Canada? At all? Then you'll also need to worry about NOM, PSE, C-Tick/A-Tick/RCM and also IC! How fun. You'll probably want this in that case:
http://www.ieee.li/pdf/essay/guide_to_global_emc_requirements_2007.pdfSo yes, in retrospect on one of the most misunderstood areas of electronics ever... there is good reason why people hate compliance laws. They are a pain in the ass and super expensive (normally). But then, so is the tooling cost of a basic plastics housing. Man, building things is really hard eh?