One possibility is to put a fixed gas frit/restrictor inline with the flow and then attach a differential pressure transducer across it and have an absolute pressure transducer at one side of the frit/restrictor.
There are equations that relate mass flow rate to the readings on these two sensors. The Freescale transducers are a reasonable price for this.
This is a technique used a lot on various instrumentation.
Temperature can affect results though so that needs to be compensated for or controlled.
Sorry - just saw note on no restriction. You might get away with a low restrictor?
How about a hot filament/thermister device? They work well for air and nitrogen. Might cost a bit more than you want. They have insignificant effect on pressure.
Try Sensirion, they were offering low cost flow sensors at one point: