diff --git a/doc/hash2/hashing_bytes.adoc b/doc/hash2/hashing_bytes.adoc
index 3a7625e..2c2f946 100644
--- a/doc/hash2/hashing_bytes.adoc
+++ b/doc/hash2/hashing_bytes.adoc
@@ -314,7 +314,7 @@ a digest of 256 bits.
 
 BLAKE2 is also designed explicitly with keyed hashing in mind, making it functionally equivalent to a Message Authentication Code (MAC). The seeded constructors act as the
 keys, BLAKE2b using a maximum of the first 64 bytes of input as the key and BLAKE2s using a maximum of the first 32 bytes. Excess input is handled by passing it to the `update()` function
-and then `result()` is called. HMAC typedefs are still provided for convenience and for compatiblity with existing protocols such as https://noiseprotocol.org/noise.html[Noise]. BLAKE2
+and then `result()` is called. HMAC typedefs are still provided for convenience and for compatibility with existing protocols such as https://noiseprotocol.org/noise.html[Noise]. BLAKE2
 behaves as all other cryptographic hash functions do when used with `hmac<H>`.
 
 ### RIPEMD-160, RIPEMD-128
diff --git a/doc/hash2/reference/sha3.adoc b/doc/hash2/reference/sha3.adoc
index 529143c..7763cb7 100644
--- a/doc/hash2/reference/sha3.adoc
+++ b/doc/hash2/reference/sha3.adoc
@@ -39,7 +39,7 @@ The only differences between them are the size of the message digest and the siz
 which is by truncating the bottom portion of the 1600-bit state.
 
 The `shake_128` and `shake_256` functions are similar to the fixed-digest functions except they use a different delimiter to denote finalization of the
-digest and extending the output has defined behavior, unlike the default pseudorandom behavior used elsewhere in the library. But they work similary
+digest and extending the output has defined behavior, unlike the default pseudorandom behavior used elsewhere in the library. But they work similarly
 in that truncations of the 1600-bit state are returned as the digest.
 
 ## sha3_256