To get mm9 FASTA files via the command-line:

$ wget http://hgdownload.cse.ucsc.edu/goldenPath/mm9/bigZips/mm9.2bit
$ wget http://hgdownload.cse.ucsc.edu/admin/exe/macOSX.x86_64/twoBitToFa
$ chmod +x ./twoBitToFa
$ for i in `seq 1 19` X Y M; do echo "converting chr$i"; ./twoBitToFa -seq=chr$i mm9.2bit chr$i.fa; done

If you are using Linux, get the twoBitToFa Kent tool with the following URL:

$ wget http://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/twoBitToFa

Install samtools. On OS X, if you have Homebrew installed, you could use brew install samtools. On Ubuntu, you might run sudo apt-get install samtools. Or on a RedHat-like Linux, you might run sudo yum install samtools.

Index the FASTA files with samtools faidx:

$ for i in `seq 1 19` X Y M; do echo "indexing chr$i"; samtools faidx chr$i.fa; done

Then query coordinates with samtools faidx. Here is a convenience Perl script I wrote that wraps around samtools, which reads stranded or unstranded BED from standard input and writes FASTA to standard output:

To use this script, e.g.:

$ ./bed2faidxsta.pl < foo.bed > foo.fa

The following link should also be helpful: Perl To Retrieve Sequences From Ucsc Genome Browser

pyfaidx has a script for this that is easy to install and works well: https://github.com/mdshw5/pyfaidx#cli-script-faidx

samtools faidx can do it too.