diff --git a/doc/dev_ref/fuzzing.rst b/doc/dev_ref/fuzzing.rst
index 21891450f62..501ad337f3f 100644
--- a/doc/dev_ref/fuzzing.rst
+++ b/doc/dev_ref/fuzzing.rst
@@ -9,16 +9,19 @@ the library.
 Fuzzing with libFuzzer
 ------------------------
 
-As of Clang Version 6.0 libFuzzer is automatically included in the compiler. Therefore you don't need to install any new software.
+libFuzzer is provided by Clang's sanitizer runtime. You do not need a separate
+libFuzzer checkout; link the fuzz targets with ``-fsanitize=fuzzer``.
 You can build the fuzzers by running ::
 
   $ ./configure.py --cc=clang --build-fuzzer=libfuzzer --enable-sanitizers=fuzzer
   $ make fuzzers
 
-The option `--enable-sanitizers=fuzzer` compiles the library for coverage-guided fuzzing.
-You can add additional sanitizers like `address`, `undefined` and `memory` or with/without
-additional information during building by either adding `--unsafe-fuzzer-mode` or `--with-debug-info`.
-The `coverage` sanitizer is not compatible with this configuration.
+The option ``--enable-sanitizers=fuzzer`` adds the required compile and link
+flags for coverage-guided fuzzing. You can add additional sanitizers like
+``address``, ``undefined`` and ``memory`` or with/without additional
+information during building by either adding ``--unsafe-fuzzer-mode`` or
+``--with-debug-info``. The ``coverage`` sanitizer is not compatible with this
+configuration.
 
 If you want to link additional libraries you can use the `--with-fuzzer-lib` option
 while configuring the build with configure.py.
@@ -34,27 +37,9 @@ with
 
 you can test the Fuzzers.
 
-Fuzzing with AFL++
---------------------
-
-Please make sure that you have installed AFL++ according to https://aflplus.plus/building/.
-The version of Clang should match the version of `afl-clang-fast++`/ `afl-clang-fast`.
-You can fuzz with AFL++ in LLVM mode (https://github.com/AFLplusplus/AFLplusplus/blob/stable/instrumentation/README.llvm.md) by running ::
-
-  $ ./configure.py --cc=clang --with-sanitizers --build-fuzzer=afl --unsafe-fuzzer-mode --cc-bin=afl-clang-fast++
-  $ make fuzzers
-
-For AFL++ in GCC mode make sure that you have `afl-g++-fast` installed.
-Otherwise follow https://github.com/AFLplusplus/AFLplusplus/blob/stable/instrumentation/README.gcc_plugin.md to build and install it.
-You can configure the build by running ::
-
-  $ ./configure.py --cc=gcc --with-sanitizers --build-fuzzer=afl --unsafe-fuzzer-mode --cc-bin=afl-g++-fast
-  $ make fuzzers
-
-The fuzzer binaries will be in `build/fuzzer`. To run them you need to
-run under `afl-fuzz`::
-
-  $ afl-fuzz -i corpus_path -o output_path ./build/fuzzer/binary
+Legacy AFL++ support remains available through ``--build-fuzzer=afl``, but it
+is not actively tested. If you use it, configure Botan with a matching AFL++
+compiler wrapper and run the resulting binary under ``afl-fuzz``.
 
 Fuzzing with TLS-Attacker
 --------------------------
diff --git a/src/fuzzer/fuzzers.h b/src/fuzzer/fuzzers.h
index 104c449cc4b..fe4b46adb07 100644
--- a/src/fuzzer/fuzzers.h
+++ b/src/fuzzer/fuzzers.h
@@ -7,9 +7,11 @@
 #ifndef BOTAN_FUZZER_DRIVER_H_
 #define BOTAN_FUZZER_DRIVER_H_
 
-#include <botan/chacha_rng.h>
 #include <botan/exceptn.h>
 #include <botan/internal/target_info.h>
+
+#include <memory>
+#include <span>
 #include <fstream>
 #include <iostream>
 #include <stdint.h>
@@ -20,6 +22,48 @@ static constexpr size_t max_fuzzer_input_size = 8192;
 
 extern void fuzz(std::span<const uint8_t> in);
 
+class Deterministic_Fuzzer_RNG final : public Botan::RandomNumberGenerator {
+   public:
+      bool accepts_input() const override { return false; }
+
+      std::string name() const override { return "Deterministic_Fuzzer_RNG"; }
+
+      void clear() override {
+         m_state = initial_state;
+         m_buffer = 0;
+         m_bytes_left = 0;
+      }
+
+      bool is_seeded() const override { return true; }
+
+   private:
+      void fill_bytes_with_input(std::span<uint8_t> output, std::span<const uint8_t> /*ignored*/) override {
+         for(auto& out_byte : output) {
+            if(m_bytes_left == 0) {
+               m_buffer = next_word();
+               m_bytes_left = sizeof(m_buffer);
+            }
+
+            out_byte = static_cast<uint8_t>(m_buffer & 0xFF);
+            m_buffer >>= 8;
+            --m_bytes_left;
+         }
+      }
+
+      uint64_t next_word() {
+         m_state ^= m_state >> 12;
+         m_state ^= m_state << 25;
+         m_state ^= m_state >> 27;
+         return m_state * 0x2545F4914F6CDD1DULL;
+      }
+
+      static constexpr uint64_t initial_state = 0x9E3779B97F4A7C15ULL;
+
+      uint64_t m_state = initial_state;
+      uint64_t m_buffer = 0;
+      size_t m_bytes_left = 0;
+};
+
 // Need to declare these before defining them;
 extern "C" int LLVMFuzzerInitialize(int* argc, char*** argv);
 extern "C" int LLVMFuzzerTestOneInput(const uint8_t in[], size_t len);
@@ -54,8 +98,7 @@ extern "C" int LLVMFuzzerTestOneInput(const uint8_t in[], size_t len) {
 // Some helpers for the fuzzer jigs
 
 inline std::shared_ptr<Botan::RandomNumberGenerator> fuzzer_rng_as_shared() {
-   static const std::shared_ptr<Botan::ChaCha_RNG> rng =
-      std::make_shared<Botan::ChaCha_RNG>(Botan::secure_vector<uint8_t>(32));
+   static const std::shared_ptr<Deterministic_Fuzzer_RNG> rng = std::make_shared<Deterministic_Fuzzer_RNG>();
    return rng;
 }