diff --git a/cbits/goldilocks.c b/cbits/goldilocks.c
index 8fcdeb5..71dc257 100644
--- a/cbits/goldilocks.c
+++ b/cbits/goldilocks.c
@@ -396,7 +396,7 @@ void goldilocks_poseidon2_permutation(uint64_t *state) {
 
 // compression function: input is two 4-element vector of field elements, 
 // and the output is a vector of 4 field elements
-void goldilocks_poseidon2_keyed_compress(uint64_t key, const uint64_t *x, const uint64_t *y, uint64_t *out) {
+void goldilocks_poseidon2_keyed_compress(const uint64_t *x, const uint64_t *y, uint64_t key, uint64_t *out) {
   uint64_t state[12];
   for(int i=0; i<4; i++) {
     state[i  ] = x[i];
@@ -411,7 +411,7 @@ void goldilocks_poseidon2_keyed_compress(uint64_t key, const uint64_t *x, const
 }
 
 void goldilocks_poseidon2_compress(const uint64_t *x, const uint64_t *y, uint64_t *out) {
-  goldilocks_poseidon2_keyed_compress(0, x, y, out);
+  goldilocks_poseidon2_keyed_compress(x, y, 0, out);
 }
 
 //------------------------------------------------------------------------------
diff --git a/cbits/goldilocks.h b/cbits/goldilocks.h
index 3bfc4ca..75dc63e 100644
--- a/cbits/goldilocks.h
+++ b/cbits/goldilocks.h
@@ -13,20 +13,11 @@ uint64_t goldilocks_sub(uint64_t x, uint64_t y);
 uint64_t goldilocks_mul(uint64_t x, uint64_t y);
 uint64_t goldilocks_mul_small(uint64_t x, uint32_t y);
 
-uint64_t goldilocks_div_by_2(uint64_t x);
-uint64_t goldilocks_div_by_3(uint64_t x);
-uint64_t goldilocks_div_by_4(uint64_t x);
-
-uint64_t goldilocks_add3(uint64_t x, uint64_t y, uint64_t z);
-
-//uint64_t goldilocks_rdc(__uint128_t x);
-
 //------------------------------------------------------------------------------
 
 void goldilocks_poseidon2_permutation(uint64_t *state);
-void goldilocks_monolith_permutation (uint64_t *state);
 
-void monolith_print_sbox_table();
-void monolith_print_sbox_table_c_format();
+void goldilocks_poseidon2_keyed_compress(const uint64_t *x, const uint64_t *y, uint64_t key, uint64_t *out);
+void goldilocks_poseidon2_compress      (const uint64_t *x, const uint64_t *y,               uint64_t *out);
 
 //------------------------------------------------------------------------------
diff --git a/poseidon2/compress.nim b/poseidon2/compress.nim
new file mode 100644
index 0000000..765c1ff
--- /dev/null
+++ b/poseidon2/compress.nim
@@ -0,0 +1,14 @@
+
+import ./types
+
+proc c_compress(a, b: var Digest, key: uint64, output: var Digest) {. header: "../cbits/goldilocks.h", importc: "goldilocks_poseidon2_keyed_compress", cdecl .}
+
+# keyed compression function
+func compress*(a, b: Digest, key: uint64 = 0) : Digest =
+  var x: Digest = a
+  var y: Digest = b
+  var output: Digest
+  c_compress(x,y,key,output)
+  return output
+
+
diff --git a/poseidon2/permutation.nim b/poseidon2/permutation.nim
new file mode 100644
index 0000000..3c7915f
--- /dev/null
+++ b/poseidon2/permutation.nim
@@ -0,0 +1,11 @@
+
+import ./types
+
+# the Poseidon2 permutation (mutable, in-place version)
+proc permInPlace*(state: var State) {. header: "../cbits/goldilocks.h", importc: "goldilocks_poseidon2_permutation", cdecl .}
+
+# the Poseidon2 permutation (pure version)
+func perm*(state: State): State =
+  var tmp = state
+  permInPlace(tmp)
+  return tmp
diff --git a/poseidon2/types.nim b/poseidon2/types.nim
index 9d3e5e5..8df0f35 100644
--- a/poseidon2/types.nim
+++ b/poseidon2/types.nim
@@ -26,4 +26,10 @@ type F12* = array[12, F]
 type Digest* = distinct F4
 type State*  = distinct F12
 
+func fromDigest* (x : Digest): F4  = return F4(x)  
+func fromState * (x : State):  F12 = return F12(x) 
+
+func toDigest* (x : F4 ): Digest = Digest(x)
+func toState*  (x : F12): State  = State(x)
+
 #-------------------------------------------------------------------------------
diff --git a/tests/poseidon2/testCompress.nim b/tests/poseidon2/testCompress.nim
new file mode 100644
index 0000000..64e75a3
--- /dev/null
+++ b/tests/poseidon2/testCompress.nim
@@ -0,0 +1,83 @@
+
+import std/unittest
+# import std/sequtils
+
+import poseidon2/types
+import poseidon2/compress
+
+#-------------------------------------------------------------------------------
+
+const refInp1: array[4,F] = 
+  [ toF( 1'u64 )
+  , toF( 2'u64 )
+  , toF( 3'u64 )
+  , toF( 4'u64 )
+  ]
+
+const refInp2: array[4,F] =  
+  [ toF( 5'u64 )
+  , toF( 6'u64 )
+  , toF( 7'u64 )
+  , toF( 8'u64 )
+  ]
+
+#---------------------------------------
+
+const refOutKey0: array[4,F] =  
+  [ toF( 0xc4a4082f411ba790'u64 )
+  , toF( 0x98c2ed7546c44cce'u64 )
+  , toF( 0xc9404f373b78c979'u64 )
+  , toF( 0x65d6b3c998920f59'u64 )
+  ]
+
+const refOutKey1: array[4,F] =  
+  [ toF( 0xca47449a05283778'u64 )
+  , toF( 0x08d3ced2020391ac'u64 )
+  , toF( 0xda461ea45670fb12'u64 )
+  , toF( 0x57f2c0b6c98a05c5'u64 )
+  ]
+
+const refOutKey2: array[4,F] =  
+  [ toF( 0xe6fcec96a7a7f4b0'u64 )
+  , toF( 0x3002a22356daa551'u64 )
+  , toF( 0x899e2c1075a45f3f'u64 )
+  , toF( 0xf07e38ccb3ade312'u64 )
+  ]
+
+const refOutKey3: array[4,F] =  
+  [ toF( 0x9930cff752b046fb'u64 )
+  , toF( 0x41570687cadcea0b'u64 )
+  , toF( 0x3ac093a5a92066c7'u64 )
+  , toF( 0xc45c75a3911cde87'u64 )
+  ]
+
+#-------------------------------------------------------------------------------
+
+suite "compression":
+
+  test "compression of [1..4] and [5..8] with key=0":
+    let input1 : Digest = toDigest(refInp1)
+    let input2 : Digest = toDigest(refInp2)
+    let output : Digest = compress(input1, input2)
+    check ( fromDigest(output) == refOutKey0 )
+
+  test "compression of [1..4] and [5..8] with key=1":
+    let input1 : Digest = toDigest(refInp1)
+    let input2 : Digest = toDigest(refInp2)
+    let output : Digest = compress(input1, input2, key=1)
+    check ( fromDigest(output) == refOutKey1 )
+
+  test "compression of [1..4] and [5..8] with key=2":
+    let input1 : Digest = toDigest(refInp1)
+    let input2 : Digest = toDigest(refInp2)
+    let output : Digest = compress(input1, input2, key=2)
+    check ( fromDigest(output) == refOutKey2 )
+
+  test "compression of [1..4] and [5..8] with key=3":
+    let input1 : Digest = toDigest(refInp1)
+    let input2 : Digest = toDigest(refInp2)
+    let output : Digest = compress(input1, input2, key=3)
+    check ( fromDigest(output) == refOutKey3 )
+
+#-------------------------------------------------------------------------------
+
diff --git a/tests/poseidon2/testPermutation.nim b/tests/poseidon2/testPermutation.nim
new file mode 100644
index 0000000..d7ba1d3
--- /dev/null
+++ b/tests/poseidon2/testPermutation.nim
@@ -0,0 +1,50 @@
+
+import std/unittest
+# import std/sequtils
+
+import poseidon2/types
+import poseidon2/permutation
+
+#-------------------------------------------------------------------------------
+
+const refInp: F12 = 
+  [ toF(  0'u64 )
+  , toF(  1'u64 )
+  , toF(  2'u64 )
+  , toF(  3'u64 )
+  , toF(  4'u64 )
+  , toF(  5'u64 )
+  , toF(  6'u64 )
+  , toF(  7'u64 )
+  , toF(  8'u64 )
+  , toF(  9'u64 )
+  , toF( 10'u64 )
+  , toF( 11'u64 )
+  ]
+
+const refOut: F12 = 
+  [ toF( 0x01eaef96bdf1c0c1'u64 )
+  , toF( 0x1f0d2cc525b2540c'u64 )
+  , toF( 0x6282c1dfe1e0358d'u64 )
+  , toF( 0xe780d721f698e1e6'u64 )
+  , toF( 0x280c0b6f753d833b'u64 )
+  , toF( 0x1b942dd5023156ab'u64 )
+  , toF( 0x43f0df3fcccb8398'u64 )
+  , toF( 0xe8e8190585489025'u64 )
+  , toF( 0x56bdbf72f77ada22'u64 )
+  , toF( 0x7911c32bf9dcd705'u64 )
+  , toF( 0xec467926508fbe67'u64 )
+  , toF( 0x6a50450ddf85a6ed'u64 )
+  ]
+
+#-------------------------------------------------------------------------------
+
+suite "permutation":
+
+  test "permutation of [0..11]":
+    var input  = toState(refInp)
+    var output = perm(input);
+    check fromState(output) == refOut
+
+#-------------------------------------------------------------------------------
+
diff --git a/tests/test.nim b/tests/test.nim
index dd76375..ec1f63f 100644
--- a/tests/test.nim
+++ b/tests/test.nim
@@ -7,5 +7,7 @@
 #import ./poseidon2/testReadme
 
 import ./poseidon2/testField
+import ./poseidon2/testPermutation
+import ./poseidon2/testCompress
 
 {.warning[UnusedImport]: off.}