people (plural for people everything is is wa: watu wetu wale wakubwa wamekuja - "those big people of ours have come")

other creatures that move around seemingly on their own initiative (animals, insects, fishes)

plants ("orange tree" is mchungwa, pl. michungwa)

edible parts of plants ("orange" is chungwa, "oranges" is machungwa)

dead things and tools we use in life bulk matter (earth, maize porridge) and abstract concepts (badness, old age)

words referring to time and place, that is where and when something is (or was, will be) to be found

It has advantages: when you hear for instance, the little boys of the neighbors say ni kidogo, "it is small", you can be sure it is not about each others penises because none of the sixteen words for penis are in the KI/VI - class.

This division in 7 classes is good enough to learn Kiswahili, though there are a lot of exceptions and even hybrid classes.

For verbs, you have a number of postfixes: funga (shut), fungwa (be shut), fungika (become shut), fungia (shut for), fungisha (cause to shut), and so on.
 

Pronunciation.

Kiswahili is easy to pronounce for westerners, especially when you speak Italian. The syllable pattern is (single) consonant-vowel-(single) consonant-vowel. The word ending is always a vowel, the stress is always on the vowel before the last one. Hence my name "Bert" is difficult: a double consonant ending. The easiest and most common Kiswahili  solution is to say Btti. Kiswahili speakers who go for pleasing me with a final consonant torture their tongues in order to end up with Brrelt. If any Dutchman called Henk does not want to be called Heki, this requires his daily attention. Kiswahili girls called Jane are always called Jani and used to that. Peter will be Pta. Some English words used by Kiswahili speakers: bar ba (two separate a's with the stress on the first one). Shirt shti. Spare spa. Bus: bsi. This prohibition of stress on the last vowel is an absolute dogma in Kiswahili. Even when that last vowel has the stress of the speakers intention, he will pronounce it unstressed. For instance je (how) in alifanyeje (how did he do it) leaves the honour of stress to ye which has little meaning apart from announcing that an item in the ("after ye")-class of endings is going to follow.
The consonant-vowel-consonant-vowel preference in Kiswahili does not only imply that often vowels are added. Sometimes they are skipped for the same reason. For instance, to Kiswahili speakers, m is a vowel in many cases, so they do not need an additional vowel o as the Portuguese in "Mozambique": Msumbiji.

A more confusing difficulty, however, is the distinction between "l" and "r". When Kiswahili people speak English, I often hear "praying" when they want to say ""playing", and "playing" when they want to say "praying". The amazing thing is that they have no plobrem understanding the westerner's English pronunciation of "l" and "r". But I experienced absurd conversations like these:

Jeremia: at the next junction, we go light .
I: What is right in Kiswahili, Jeremia?
Jeremia: Kuria
I: And left?
Jeremia: Kushoto.
I: OK, so I can easily remember it: kuria has the "r" of "right"...
Jeremia: Yes, the r of light.

The dictionary turns out to give for "right": kulia with the l of "left".

Different Bantu languages generate "r" resp. "l" impressions in western ears at different praces. There is no genelar lure.
Even Kiswahili text books find it hard to maintain "l" and "r" on there own praces in the arfabet.

Vocabulary

Tu is not "you", as in Roman languages, but "we"
M is not "me" but "you"(plural), and also "him", and thirdly, it can denote present perfect, culminating in expressions like (Mmempa yeye You have given him − M-me-m-pa You-have-him-give)
Nane looks like "nine" in many indo European languages but it is "eight"
Sita is "six", and should not be confused with Tisa
Tisa looks like "thousand" but it is "nine"
Mbili looks like "mille" or "million" thus could thought to be "thousand" or "million", but it is "two".
Elfu is not "eleven" but "thousand".
Si, "yes" in Roman languages, means "is not", negation
Ni, negation in all Indo-Germanic languages, is "yes", affirming, positive.
Dada does not mean "father" but "sister"
Amekaa mbele fortunately does not mean "American car", it means: "the one sitting in front"
Hawana unfortunately has nothing to to with my cherished Cuban cigars, that are very hard to get here. Hawana Havana means: "they have no Havana's".

Kiswahili has arisen out of the necessity for Bantu tribes to communicate in the East African trade boosted by the Arabs. It contains a host of Arab words. Arabs had a sophisticated money economy (check is an Arab invention and an Arab word!). It is all the more surprising that in Kiswahili, "bank" is benk, "change" is chenji. For "money" itself there are three words, fedha (Arab), hela (Hindi, Indians are the backbone of the East African economy) and pesa, the origin of which, according to the authoritative Johnson dictionary, is the Hindi "". That makes us end up without any traditional Bantu word for money.

Quite thoughtfully Tumo baa ("We are in the bar"), has been chosen such as to be understandable even if spoken after quite some beers through a bad phone line. Why Kiswahili waited for the English to provide a word for "blue", (buluu), is an intriguing question. Kiswahili people tend not to give names to colors, but to say "the colour of...." (leaf, dry leaf, coffee, ash, turmeric)

The Germans and Dutch will be pleased to note that a "small mountain" (mlima) is kilima, a "small child" (mtoto) is kitoto etc. (so the German and Dutch klein, and not the English "small" or French petit). Mengi is many (Dutch: menige), ingine can mean "few" (German einige, Dutch: enige). "To enter" is ingia, in Dutch: ingaan. "Egg" is yai (Dutch ei, pronounces ai in German, including some Dutch dialects). All purely accidental, or would it have a background in instinctive primitive sound language developed in the time our ancestors were still laying eggs? "Ai!!!" Ultimate satisfaction for Germans: "School" is Schule in Kiswahili, but this really is a German loan-word.

To my utter astonishment Kiswahili has a word for a unit of seven days, hence "week". Why seven? How on earth would non Mesopotamic peoples have such a 7 days unit? It it is not based on any astronomical cycle whatsoever hence completely arbitrary. Answer: from de Muslims: "day 1" is Saturday, "day 7" is Friday. This leaves us in limbo concerning the true traditional Bantu handling of days, months and years.

Dictionary problems

Words, both nouns and verbs, get prefixes. The stem ti means chair, but one always hears or sees it as kiti (sg.) or viti (pl.). Pend is "want", but you always see or hear it like ninapenda, "I want", sipendi, "I do not want" (i is a postfix and there are many types). Hence one should know the word stem to find a word in an alphabetical dictionary, typically something you do not know if you need to search for its meaning.
I do not know for sure, but I have little doubt that the first Kiswahili writing was in Arab letters. East Africa was at the fringe of Arab culture for at least nine centuries before it was "discovered" (as the retarded Christians call the event of their own first arrival after everyone else). No doubt reading Kiswahili in Arab letters is much more appropriate, not only since you read in the opposite direction (somebody told me I make a mistake here, but his argument was to complicated for me).