I have been pleased with this unusual AGC circuit as it nicely removes the top of large signals. VK3YE has suggested measuring current in parallel with the LED that drives the AGC action, but I found this gave little indication of relative signal strength. it works much better measuring current in series with the LED.
I am using a small meter movement supplied by Sunil, about 250 uV peak current. A shunt resistor across the meter is necessary since the LED current at peak is at least 20 mA per my research. The approximate value of the shunt resistor Rsh = ( Im x Rm ) / Ish
Rm was unknown but I am able to measure it with my DVM at around 500 ohms. Inserting these values gives 6.25 ohms. i found a resistor around 5 ohms and it working nicely. Its not a real S-meter response - let's call it a signal strength meter. I sense I can discern signals from approx S5 to S9 with it. Now if someone does not have a meter movement, the LED itself could be observed to sense signal strength - but the meter is a nice touch.
I have previously tried s-meter sensing circuits but the action of this particular AGC circuit prohibits them from functioning - so this simple current meter does decently.