SayMore is probably too heavyweight for what you're doing. One nice way to gather words and recordings in a more light-weight way would be WeSay (our simple dictionary program) or FLEx (our powerful dictionary program for linguists). Both let you type in a word, a gloss, and click the "record" button. I think both then save the audio to a .wav file you can get at on your disk. You could then use those in TRP somehow.
When I studied Thai and did TPR, I recorded with WeSay also added images for each word. Then I wrote a little program called Subachup that would show me a grid of images, play the recording, and let me click on whichever image I thought matched what it heard. It tracked how often I was getting them right and wrong, and would cycle back to the ones I needed work on. In that way, it eventually zeroed in on minimal pairs that had sounds I didn't hear well (e.g. vowel length). I loved the visual and game-like nature of it.
Alas, I never got that out into the real world, so it's of no use to you. But I'm describing it in case that sounds like something you'd like, well now you can go looking to see if anyone else had the same idea and published software like that.
Another approach that is definetly shipping is Jonathan Coomb's FlashGrab
, which can actually import words from the .LIFT file format that both WeSay and FLEx support.
Hope that helps!