diff --git a/src/TextAnalysis.jl b/src/TextAnalysis.jl
index e34285b0..cdc761b3 100644
--- a/src/TextAnalysis.jl
+++ b/src/TextAnalysis.jl
@@ -6,6 +6,7 @@ module TextAnalysis
     using Languages
     using DataFrames
     using WordTokenizers
+    using DataStructures
 
     import DataFrames.DataFrame
     import Base.depwarn
@@ -45,6 +46,7 @@ module TextAnalysis
     export tf, tf_idf, lsa, lda, summarize
     export tf!, tf_idf!, lsa!, lda!
     export remove_patterns!, remove_patterns
+    export fastpreprocess, PreprocessBuffer
 
     export strip_patterns, strip_corrupt_utf8, strip_case, stem_words, tag_part_of_speech, strip_whitespace, strip_punctuation
     export strip_numbers, strip_non_letters, strip_indefinite_articles, strip_definite_articles, strip_articles
@@ -61,6 +63,9 @@ module TextAnalysis
     include("corpus.jl")
     include("metadata.jl")
     include("preprocessing.jl")
+
+    include("fastpreprocess.jl")
+
     # Load libstemmer from our deps.jl
     const depsjl_path = joinpath(dirname(@__FILE__), "..", "deps", "deps.jl")
     if !isfile(depsjl_path)
diff --git a/src/fastpreprocess.jl b/src/fastpreprocess.jl
new file mode 100644
index 00000000..faf97697
--- /dev/null
+++ b/src/fastpreprocess.jl
@@ -0,0 +1,246 @@
+# TODO Figure out the following:
+# * strip_sparse_terms - to utilize `words_remove` and `sparse_terms` (of preprocessing.jl).
+# * strip_frequent_terms - to utilize `words_remove` and `frequent_terms` (of preprocessing.jl).
+# * strip_html_tags
+# * strip_non_letters
+# * strip_case
+"""
+Preprocessing functions
+
+* corrupt_utf8
+* whitespace
+* punctuation
+* numbers
+* indefinite_articles
+* definite_articles
+* articles
+* stopwords
+* prepositions
+* pronouns
+
+
+Turns a string into a readable and writable stream of `Char`s,
+used for preprocessing and flushing out the processes text.
+
+Utility functions (lexers) such as `spaces` and `number` read characters from the stream
+and match against it.
+
+Functions (lexers) return `true` or `false` to indicate whether they matched anything
+in the input stream. They can therefore be combined easily, e.g.
+
+    spacesornumber(ts) = whtiespace(ts) || numbers(ts)
+
+either deletes two consectutively read whitespaces or removes a number character, if matched.
+
+For certain cases like `strip_pronouns`, `strip_prepositions`, `strip_stopwords`, etc.
+These are stored into a `SortedSet` for faster preprocessing and
+matches words / tokens against the characters in the stream
+in the function `words_remove`.
+"""
+mutable struct PreprocessBuffer
+    input::Vector{Char}
+    idx::Int
+end
+
+PreprocessBuffer(input) = PreprocessBuffer(input, 1)
+
+PreprocessBuffer(input::AbstractString) = PreprocessBuffer(collect(input))
+
+Base.getindex(ps::PreprocessBuffer, i = ps.idx) = ps.input[i]
+
+isdone(ps::PreprocessBuffer) = ps.idx > length(ps.input)
+
+# TODO: Remove whitespace at the end, beginning and multiple whitepsaces into one.
+"""
+    corrupt_utf8(ps::PreprocessBuffer)
+
+Removes the corrupt UTF8 chars.
+"""
+function corrupt_utf8(ps)
+    isvalid(ps[ps.idx]) && return false
+
+    deleteat!(ps.input, ps.idx)
+    return true
+end
+
+"""
+    whitespace(ps::PreprocessBuffer)
+
+Squash multiple whitespaces to a single one.
+And remove all leading and trailing whitespaces.
+"""
+function whitespace(ps)
+    isspace(ps[ps.idx]) || return false
+
+    ps.idx != 1 && !isspace(ps[ps.idx - 1]) && return next(ps)
+
+    deleteat!(ps.input, ps.idx)
+    return true
+
+    # If prev is whitespace then delete.
+end
+
+"""
+    trailing_whitespace(ps::PreprocessBuffer)
+
+Remove the whitespaces at the end of the input stream.
+"""
+function trailing_whitespace(ps)
+    isspace(ps[length(ps.input)]) || return
+    i = length(ps.input) - 1
+
+    while (i > 0) && isspace(ps[i])
+        i -= 1
+    end
+
+    deleteat!(ps.input, i + 1: length(ps.input))
+end
+
+"""
+    punctuation(ps::PreprocessBuffer)
+
+Remove punctuations, as matched by `ispunct`.
+"""
+function punctuation(ps)
+    ispunct(ps[]) || return false
+
+    deleteat!(ps.input, ps.idx)
+    return true
+end
+
+"""
+    numbers(::PreprocessBuffer)
+
+Removes all numbers.
+"""
+function numbers(ps)
+    isdigit(ps[]) || return false
+
+    deleteat!(ps.input, ps.idx)
+    return true
+end
+
+"""
+Helper function for words_remove.
+Matches the next token in the stream against the `ws::SortedSet`.
+Returns whether it matched and the idx of the token end
+"""
+function next_token(ps::PreprocessBuffer, ws)
+    i = ps.idx
+    while i <= length(ps.input) && isletter(ps[i])
+        i += 1
+    end
+    i < length(ps.input) && isdigit(ps[i]) && return false, i
+
+    String(ps.input[ps.idx:i-1]) ∈ ws && return true, i
+    return false, i
+end
+
+"""
+Matches true for characters corresponding to Regex("[a-zA-Z0-9_]")
+"""
+word_character(ch) = isascii(ch) && (isuppercase(ch) || islowercase(ch) ||
+                                            isdigit(ch) || ch == '_')
+
+"""
+    words_remove(::PreprocessBuffer, ws)
+
+Removes words from the PreprocessBuffer.
+"""
+function words_remove(ps, ws)
+    ps.idx != 1 && word_character(ps[ps.idx - 1]) && return false
+    isletter(ps[ps.idx]) || return false
+
+    match, i = next_token(ps, ws)
+
+    if match == false
+        ps.idx = i
+    else
+        deleteat!(ps.input, ps.idx:i - 1)
+    end
+
+    return true
+end
+
+function next(ps::PreprocessBuffer)
+    ps.idx += 1
+    return true
+end
+
+"""
+    fastpreprocess(::StringDocument, flags)
+    fastpreprocess(::Corpus, flags)
+    fastpreprocess(::String, lang::T, flags) where T <: Language
+    fastpreprocess(::String, ::SortedSet, flags)
+
+## Preprocessing functions currently available
+
+* corrupt_utf8
+* whitespace
+* punctuation
+* numbers
+
+### Flags for functions requiring `words_remove`
+
+* strip_indefinite_articles
+* strip_definite_articles
+* strip_articles
+* strip_stopwords
+* strip_prepositions
+* strip_pronouns
+
+## Usage
+
+
+## Note:
+
+This does not work for Corpora consisting of `FileDocument`,
+`TokenDocument` or `NGramDocument`
+
+"""
+fastpreprocess(txt::String, lang = Languages.English(), flags = 0) = fastpreprocess(txt, build_set(flags, lang))
+
+function build_set(flags, lang = Languages.English())
+    ws = SortedSet()
+
+    ((flags & strip_indefinite_articles) > 0) && union!(ws, indefinite_articles(lang))
+    ((flags & strip_definite_articles) > 0) && union!(ws, definite_articles(lang))
+
+    ((flags & strip_prepositions) > 0) && union!(ws, prepositions(lang))
+    ((flags & strip_pronouns) > 0) && union!(ws, pronouns(lang))
+    ((flags & strip_stopwords) > 0) && union!(ws, stopwords(lang))
+    ws
+end
+
+# TODO: Check case insensitive in words
+function fastpreprocess(txt::String, ws::SortedSet)
+    length(txt) < 1 && return
+    ps = PreprocessBuffer(txt)
+
+    while !isdone(ps)
+        whitespace(ps) ||
+        corrupt_utf8(ps) ||
+        punctuation(ps) ||
+        numbers(ps) ||
+        words_remove(ps, ws) || next(ps)
+    end
+
+    trailing_whitespace(ps)
+    return String(ps.input)
+end
+
+function fastpreprocess(doc::StringDocument, flags = 0)
+    doc.text =  fastpreprocess(doc.text, build_set(flags, language(doc)))
+end
+
+# Only for String Document
+function fastpreprocess(crps::Corpus, flags = 0)
+    ws = build_set(flags, language(crps[1]))
+
+    for doc in crps
+        doc.text = fastpreprocess(doc.text, ws)
+    end
+    crps
+end
+
+# HTML placed before words
diff --git a/test/fastpreprocess.jl b/test/fastpreprocess.jl
new file mode 100644
index 00000000..c174a661
--- /dev/null
+++ b/test/fastpreprocess.jl
@@ -0,0 +1,36 @@
+@testset "Preprocessing" begin
+    @testset "Words Removal" begin
+        doc = StringDocument("this is a the sample text")
+        fastpreprocess(doc, strip_articles)
+        @test isequal(doc.text, "this is sample text")
+
+        doc = Document("this is the sample text")
+        fastpreprocess(doc, strip_definite_articles)
+        @test isequal(doc.text, "this is sample text")
+
+        doc = Document("this is a sample text")
+        fastpreprocess(doc, strip_indefinite_articles)
+        @test isequal(doc.text, "this is sample text")
+
+        doc = Document("this is on sample text")
+        fastpreprocess(doc, strip_prepositions)
+        @test isequal(doc.text, "this is sample text")
+
+        doc = Document("this is my sample text")
+        fastpreprocess(doc, strip_pronouns)
+        @test isequal(doc.text, "this is sample text")
+
+        doc = Document("this is sample text")
+        fastpreprocess(doc, strip_stopwords)
+        @test isequal(strip(doc.text), "sample text")
+    end
+
+    # test Remove Corrupt UT8
+    sd = StringDocument("abc")
+    fastpreprocess(sd)
+    @test sd.text == "abc"
+
+    sd = StringDocument(String([0x43, 0xf0]))
+    fastpreprocess(sd)
+    @test sd.text == "C"
+end
diff --git a/test/runtests.jl b/test/runtests.jl
index bffc62d2..c251bdbd 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -15,6 +15,7 @@ include("document.jl")
 include("metadata.jl")
 include("corpus.jl")
 include("preprocessing.jl")
+include("fastpreprocess.jl")
 include("dtm.jl")
 include("stemmer.jl")
 include("tf_idf.jl")