meta-scrubber provides a streaming metadata remover
It is a WORK IN PROGRESS and currently provides ZERO guarantees and VERY limited file format support.
$ go build ./cmd/meta-scrubber
$ ./meta-scrubber input-file.png output-file.png
metascrubber will run tests on any .jpg or .png images in testdata.
There's a large (currently ~1GB) corpus for testing at https://meta-scrubber-test-corpus.s3.us-west-1.amazonaws.com/exif-image-corpus.tar.gz. This corpus will be automatically downloaded and cached in testdata when the tests are first run. On future test runs, the corpus is checked for changes and, if necessary, updated. To avoid downloading/updating the corpus, or just to avoid testing against the corpus, run the tests with the -short flag.