I think there is a shorter way of doing this but I use the below:

library(Rsamtools)
library(BSgenome)

# read data
dat <- read.table(text = 'chr start
                  chr1  114242
                  chr1  114242
                  chr2  7485484', header=T,stringsAsFactors=F)

# if you have a genome fasta file you can import it using FaFile
# the format of chromosomes should match that in the fasta file
# i.e. 1, 2, 3 in both fasta and dat or chr1, chr2, chr3 in both
fasta_file <- FaFile(file='hg19.fasta')
gr1 <- GRanges(dat$chr,IRanges(start=as.numeric(dat$start), end=as.numeric(dat$start)))
refbase <- getSeq(fasta_file, gr1)
refbase <- as.data.frame(refbase)$x
dat$REF <- refbase

Something like:

samtools faidx hg19.fasta "1:20000-20000"
samtools faidx hg19.fasta `tail -n +2 test.txt | awk '{ print $1":"$2"-"$2 }'`

I don't know if it works with stdin. You can explore more.

See GenomicRanges. Its documentation is really helpful.