diff --git a/README.md b/README.md
index 2c5d4af9..1e719eb5 100644
--- a/README.md
+++ b/README.md
@@ -31,7 +31,7 @@ information from the target.
 By applying fuzzing techniques to a proxy, we can generate a series of test
 inputs (corpus) that produce interesting behavior in the proxy. Our underlying
 assumption is that this translates into similarly interesting behavior in the
-target.
+target. Read more in the [docs](https://github.com/google/silifuzz/blob/main/doc/proxy_architecture.md).
 
 ##### Corpus / Corpus shard
 
@@ -50,7 +50,8 @@ microseconds, but it can be arbitrarily large. Snapshots are stored as
 
 Snapshots are typically created from the inputs generated by a fuzzing engine.
 For CPU testing purposes these inputs are filtered to eliminate all
-non-deterministic snapshots.
+non-deterministic snapshots. Read more in the
+[docs](https://github.com/google/silifuzz/blob/main/doc/what_makes_a_good_test.md).
 
 ###### Expected end state
 
@@ -67,7 +68,7 @@ in-memory one except that native pointers are replaced with offsets. See this
 [header](https://github.com/google/silifuzz/blob/main/snap/gen/relocatable_snap_generator.h)
 for details. This format is often referred to as *relocatable*. Each Snap
 contains exactly one expected end state i.e. Snaps are
-microarchitecture-specific.
+microarchitecture-specific. Read more in the [docs](https://github.com/google/silifuzz/blob/main/doc/snap.md).
 
 ##### Runner
 
diff --git a/doc/snap.md b/doc/snap.md
index 1d104a67..cc015452 100644
--- a/doc/snap.md
+++ b/doc/snap.md
@@ -1,6 +1,4 @@
-# Silifuzz Snap Data Structurue
-
-[TOC]
+# Silifuzz Snap Data Structure
 
 ## Overview