Wow ,
I really will try to make this as simple as possible.
I'm including a TLDR version below.
First we have to establish a few definitions.
Voltage Standing Wave Ratio , otherwise VSWR, vswr, or swr (yuck) This last is a bit lazy and not quite correct so I ...
Additional notes and links:
Standing up for standing waves. Probably the best 90min you'll ever spend understanding radio waves in general.
Vizwarz, or how to get it wrong on so many levels.
https://forums.qrz.com/index.php?threads/vizwarz-or-only-your-barista-knows-for-sure.594793/
Another look at reflections. This is admittedly advanced material and will either make your head hurt and eyes bleed, or put you to sleep. In any case it had to be mentioned.
http://www.k3emd.com/downloads/Reflect.pdf
Resonance.
Resonance is not impedance. Let's try that again, with feeling
, Resonance is Not VSWR. Resonance can be at minimal vswr, or not. Minimum vswr can be at resonance, or not. The two are interrelated but mutually exclusive and separate terms.
We currently have cool toys like rig expert and nano vna. but unless you anticipate frequently tossing wire around the property even the $99 VNA is overkill.
So with those definitions and links in place:
Place a dummy load on the output of the amplifier and without plugging it in or turning it on check the vswr between the radio and the amp. This first easy step will prevent issues later. Any bypass vswr has to be eliminated. Methods are a swamping resistor(s), constant impedance attenuator (3Db "T" pad) , or a lump constant tuned circuit ( capacitor/inductor network) Once the vswr barely if at all registers , move on .
Next, turn the amp on and with reasonable input observe the input vswr with the amp output into a dummy load. Again if this circuit displays significant vswr a swamping resistor(s), constant impedance attenuator (3Db "T" pad) , or a lump constant tuned circuit ( capacitor/inductor network) are needed to adjust the input vswr across the intended range of operation. I don't mean to gloss over all that this entails because you simply asked "where does the vswr meter go" . What you didn't know is that this opens a YUUUUUge can of worms.
So now you have a radio that drives an amplifier without seeing significant vswr. You have some significant output at the exit of the amplifier. What next?
A brief look at drive and efficiency.
One watt in = 10 watts out
Two watts in = 20 watts out
Three watts in = 30 watts out
Four watts in = 40 watts out
....
This is "linear" behavior.
14.49A @ 13.8V for 100 watts output.
This amplifier is 50% efficient. Take my word for it or do the math
Fully modulated AM output is four times the quiescent carrier. So four watts carrier fully modulated is 16 watts.
Now do you understand why a pair of 2sc455's or even 2sc2290's only need 2-3 watts of drive?
So the easiest way to establish drive level is watts input/output vs current draw. Simply put if the next amp you draw is only producing 25% more output than the last amp you drew it's time to back off. How much you back off depends on how deep your pockets are.
Okay now we're at the voodoo where resonance does not equal resistance. Resonance is NOT necessarily at 1:1 vswr.
A dummy load has a perfect 1:1 match but isn't much of antenna by design. Off Center Fed Dipoles are regularly fed at a 200 ohm point and are great antennas. Several constructions of wire antennas are resonant at frequency and exhibit feed point impedance in the several thousands of ohms.
From the above link to the antenna handbook find the chapter on matching sections. If the dimensions of the antenna make any realistic sense IE: 9 foot whip, 18 foot dipole, or any reasonable fraction or multiple there of you can continue.
You have two choices here, fool the amp into seeing a flat match with a tuner or fix the antenna feed point. The alternative is to place the tuner at the antenna feed point but not always convenient.
The best of the two choices as to fix the feed point so that your coax or other feed line isn't confused and either radiating or heating.
Mismatches under 2.5:1 (opinion) result in such minimal losses that placing the tuner at the amp output is acceptable if not ideal.
so the TLDR version as promised. An even shorter version follows this one.
Both goesinta and goesouta vswr/power meters.
Resonance does not equal vswr
For solid state devices an amp meter really really really helps set appropriate drive levels.
Do not over drive the poor beastie. You will let the smoke out.
If you're getting 100 ish (peak reading)output for around 15A @ 13.8V, out for 3 watts in on AM with a tone or whistle with a 25-30 dead carrier out of a random pair of magic smoke containment devices be happy.
12-14V DC MOSFET amplifiers are an absolute minefield.
I avoid them like my future ex-wife.
Amp bypass and amp input vswr are two separate circuits.
If the amp and the antenna are at or near 50 ohms the feed line becomes practically invisible save for it's losses.
It's not rocket surgery but it is brain science.
TLDR:
Both
Don't overdrive it or gawd will kill a kitten.