diff --git a/README.md b/README.md
new file mode 100644
index 0000000..4f2e5b8
--- /dev/null
+++ b/README.md
@@ -0,0 +1,10 @@
+# Upkr
+
+Upkr is a simple general purpose lz packer designed to be used in the [MicroW8](https://github.com/exoticorn/microw8) platform.
+The compressed format is base on [Shrinkler](https://github.com/askeksa/Shrinkler) with the main difference being that
+Upkr doesn't differnetiate between literals at odd or even addresses.
+
+At this point, Upkr should be considered unstable - the exact format isn't finalized yet and still subject to change
+and only a very simple (but also very fast) greedy compressor is implemented. The compression ratio will be improved
+with a more thourough lz parse in the future, although even in the current state is is already similar to the
+DEFLATE compression algorithm.
\ No newline at end of file
diff --git a/src/lib.rs b/src/lib.rs
new file mode 100644
index 0000000..cdfe38c
--- /dev/null
+++ b/src/lib.rs
@@ -0,0 +1,8 @@
+mod context_state;
+mod greedy_packer;
+mod lz;
+mod match_finder;
+mod range_coder;
+
+pub use greedy_packer::pack;
+pub use lz::unpack;
\ No newline at end of file
diff --git a/src/lz.rs b/src/lz.rs
index f220a28..9c10439 100644
--- a/src/lz.rs
+++ b/src/lz.rs
@@ -53,7 +53,7 @@ impl LzCoder {
         assert!(value >= 1);
         let top_bit = usize::BITS - 1 - value.leading_zeros();
         let mut context_index = context_start;
-        for i in (0..top_bit).rev() {
+        for i in 0..top_bit {
             self.bit(true, context_index);
             self.bit((value >> i) & 1 != 0, context_index + 1);
             context_index += 2;
@@ -78,14 +78,16 @@ pub fn unpack(packed_data: &[u8]) -> Vec<u8> {
         contexts: &mut ContextState,
         mut context_index: usize,
     ) -> usize {
-        let mut length = 1;
+        let mut length = 0;
+        let mut bit_pos = 0;
         while decoder.decode_with_context(&mut contexts.context_mut(context_index)) {
-            length = (length << 1)
-                | decoder.decode_with_context(&mut contexts.context_mut(context_index + 1))
-                    as usize;
+            length |= (decoder.decode_with_context(&mut contexts.context_mut(context_index + 1))
+                as usize)
+                << bit_pos;
+            bit_pos += 1;
             context_index += 2;
         }
-        length
+        length | (1 << bit_pos)
     }
 
     loop {
diff --git a/src/main.rs b/src/main.rs
index 82705bb..d3a372b 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,16 +1,10 @@
-mod context_state;
-mod greedy_packer;
-mod lz;
-mod match_finder;
-mod range_coder;
-
 fn main() {
-    let test_data = include_bytes!("../testcases/skipahead.wasm");
+    let test_data = include_bytes!("../README.md");
 
-    let packed = greedy_packer::pack(test_data);
+    let packed = upkr::pack(test_data);
     dbg!((test_data.len(), packed.len()));
 
-    let unpacked = lz::unpack(&packed);
+    let unpacked = upkr::unpack(&packed);
     dbg!(unpacked.len());
     assert!(test_data == unpacked.as_slice());
 }
diff --git a/src/match_finder.rs b/src/match_finder.rs
index e576019..ae70fd7 100644
--- a/src/match_finder.rs
+++ b/src/match_finder.rs
@@ -113,7 +113,7 @@ impl<'a> Iterator for Matches<'a> {
                     self.add_to_queue(self.finder.suffixes[self.left_index]);
                     self.move_left();
                 }
-                if self.right_length == self.current_length {
+                if self.right_length == self.current_length && self.matches_left > 0 {
                     self.add_to_queue(self.finder.suffixes[self.right_index]);
                     self.move_right();
                 }