We've transitioned our local/dev/prod instances to use conda on Heroku, and couldn't be happier. It was a tiny bit of work to get it set up, but now everything is consistent, and we can set up new local environments in seconds.
I really only use conda for the non-python bits of our stack: numpy/scipy/pandas etc - packages that are a pain to install on Heroku.
Especially with all the blas linking. Was there anything special you had to do or was it simply conda install numpy-blas or something like that ?
We use conda-env (https://github.com/conda/conda-env) which makes it really simple to manage environments locally and in remote environments.
Incidentally is the buildscript of anaconda itself opensource? Couldn't find it anywhere.
Instead of requirements.txt we use py.prereqs, but otherwise this script should get you close.