diff --git a/Cargo.toml b/Cargo.toml
index 37c294b..77b65f2 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -12,10 +12,12 @@ documentation = "https://docs.rs/instant-segment"
 [features]
 __test_data = ["test-cases"]
 test-cases = []
+with-serde = ["serde", "ahash/serde", "smartstring/serde"]
 
 [dependencies]
 ahash = "0.7.0"
 smartstring = "0.2.5"
+serde = { version = "1.0.123", features = ["derive"], optional = true }
 
 [dev-dependencies]
 bencher = "0.1.5"
diff --git a/src/lib.rs b/src/lib.rs
index 73d86f4..df22809 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -1,6 +1,8 @@
 use std::ops::{Index, Range};
 use std::str;
 
+#[cfg(feature = "serde")]
+use serde::{Deserialize, Serialize};
 use smartstring::alias::String;
 
 #[cfg(feature = "test-cases")]
@@ -9,6 +11,7 @@ pub mod test_cases;
 pub mod test_data;
 
 /// Central data structure used to calculate word probabilities
+#[cfg_attr(feature = "with-serde", derive(Deserialize, Serialize))]
 pub struct Segmenter {
     unigrams: HashMap<String, f64>,
     bigrams: HashMap<(String, String), f64>,