diff --git a/fe25519_25.js b/fe25519_25.js
index 0349979..8f5bd8a 100644
--- a/fe25519_25.js
+++ b/fe25519_25.js
@@ -35,6 +35,7 @@ const importWithMemory = {
 }
 
 const wasm_mul = require('./fe25519_25/fe25519_mul')(importWithMemory)
+const wasm_mul32 = require('./fe25519_25/fe25519_mul32')()
 const wasm_sq = require('./fe25519_25/fe25519_sq')(importWithMemory)
 const wasm_invert = require('./fe25519_25/fe25519_invert')(importWithMemory)
 const wasm_pow = require('./fe25519_25/fe25519_pow22523')()
@@ -799,6 +800,22 @@ function fe25519_mul (h, f, g) {
   parse_fe(h, mem, 80)
 }
 
+function fe25519_mul32 (h, f, n) {
+  check_fe(h)
+  check_fe(f)
+
+  // printFe(f, 'f')
+  // printFe(g, 'g')
+
+  var fbuf = new Uint8Array(f.buffer)
+
+  wasm_mul32.memory.set(fbuf)
+  wasm_mul32.exports.fe25519_mul32(40, 0, n)
+
+  const output = Buffer.from(wasm_mul32.memory.slice(40, 80))
+  parse_fe(h, output, 0)
+}
+
 /*
  h = f * f
  Can overlap h with f.
diff --git a/fe25519_25/fe25519_mul32.js b/fe25519_25/fe25519_mul32.js
new file mode 100644
index 0000000..c9625d5
--- /dev/null
+++ b/fe25519_25/fe25519_mul32.js
@@ -0,0 +1,61 @@
+
+module.exports = loadWebAssembly
+
+loadWebAssembly.supported = typeof WebAssembly !== 'undefined'
+
+function loadWebAssembly (opts) {
+  if (!loadWebAssembly.supported) return null
+
+  var imp = opts && opts.imports
+  var wasm = toUint8Array('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')
+  var ready = null
+
+  var mod = {
+    buffer: wasm,
+    memory: null,
+    exports: null,
+    realloc: realloc,
+    onload: onload
+  }
+
+  onload(function () {})
+
+  return mod
+
+  function realloc (size) {
+    mod.exports.memory.grow(Math.max(0, Math.ceil(Math.abs(size - mod.memory.length) / 65536)))
+    mod.memory = new Uint8Array(mod.exports.memory.buffer)
+  }
+
+  function onload (cb) {
+    if (mod.exports) return cb()
+
+    if (ready) {
+      ready.then(cb.bind(null, null)).catch(cb)
+      return
+    }
+
+    try {
+      if (opts && opts.async) throw new Error('async')
+      setup({instance: new WebAssembly.Instance(new WebAssembly.Module(wasm), imp)})
+    } catch (err) {
+      ready = WebAssembly.instantiate(wasm, imp).then(setup)
+    }
+
+    onload(cb)
+  }
+
+  function setup (w) {
+    mod.exports = w.instance.exports
+    mod.memory = mod.exports.memory && mod.exports.memory.buffer && new Uint8Array(mod.exports.memory.buffer)
+  }
+}
+
+function toUint8Array (s) {
+  if (typeof atob === 'function') return new Uint8Array(atob(s).split('').map(charCodeAt))
+  return (require('buf' + 'fer').Buffer).from(s, 'base64')
+}
+
+function charCodeAt (c) {
+  return c.charCodeAt(0)
+}
diff --git a/fe25519_25/fe25519_mul32.wat b/fe25519_25/fe25519_mul32.wat
new file mode 100644
index 0000000..0e28720
--- /dev/null
+++ b/fe25519_25/fe25519_mul32.wat
@@ -0,0 +1,161 @@
+(module
+  (memory $0 1)
+  (export "memory" (memory $0))
+
+  ;; (func $i32.log (import "debug" "log") (param i32))
+  ;; (func $i32.log_tee (import "debug" "log_tee") (param i32) (result i32))
+  ;; ;; No i64 interop with JS yet - but maybe coming with WebAssembly BigInt
+  ;; ;; So we can instead fake this by splitting the i64 into two i32 limbs,
+  ;; ;; however these are WASM functions using i32x2.log:
+  ;; (func $i32x2.log (import "debug" "log") (param i32) (param i32))
+  ;; (func $f32.log (import "debug" "log") (param f32))
+  ;; (func $f32.log_tee (import "debug" "log_tee") (param f32) (result f32))
+  ;; (func $f64.log (import "debug" "log") (param f64))
+  ;; (func $f64.log_tee (import "debug" "log_tee") (param f64) (result f64))  
+  
+  ;; ;; i64 logging by splitting into two i32 limbs
+  ;; (func $i64.log
+  ;;   (param $0 i64)
+  ;;   (call $i32x2.log
+  ;;     ;; Upper limb
+  ;;     (i32.wrap/i64
+  ;;       (i64.shr_s (get_local $0)
+  ;;         (i64.const 32)))
+  ;;     ;; Lower limb
+  ;;     (i32.wrap/i64 (get_local $0))))
+
+  ;; (func $i64.log_tee
+  ;;   (param $0 i64)
+  ;;   (result i64)
+  ;;   (call $i64.log (get_local $0))
+  ;;   (return (get_local $0)))
+
+  (func $fe_mul32
+    (param $f0 i64)
+    (param $f1 i64)
+    (param $f2 i64)
+    (param $f3 i64)
+    (param $f4 i64)
+    (param $f5 i64)
+    (param $f6 i64)
+    (param $f7 i64)
+    (param $f8 i64)
+    (param $f9 i64)
+
+    (param $n i32)
+
+    (param $h i32)
+
+    (local $sn i64)
+
+    (local $h0 i64)
+    (local $h1 i64)
+    (local $h2 i64)
+    (local $h3 i64)
+    (local $h4 i64)
+    (local $h5 i64)
+    (local $h6 i64)
+    (local $h7 i64)
+    (local $h8 i64)
+    (local $h9 i64)
+    
+    (local $carry0 i64)
+    (local $carry1 i64)
+    (local $carry2 i64)
+    (local $carry3 i64)
+    (local $carry4 i64)
+    (local $carry5 i64)
+    (local $carry6 i64)
+    (local $carry7 i64)
+    (local $carry8 i64)
+    (local $carry9 i64)
+
+    (set_local $sn (i64.extend_s/i32 (get_local $n)))
+
+    (set_local $f0 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f0))))
+    (set_local $f1 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f1))))
+    (set_local $f2 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f2))))
+    (set_local $f3 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f3))))
+    (set_local $f4 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f4))))
+    (set_local $f5 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f5))))
+    (set_local $f6 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f6))))
+    (set_local $f7 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f7))))
+    (set_local $f8 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f8))))
+    (set_local $f9 (i64.extend_s/i32 (i32.wrap/i64 (get_local $f9))))
+
+    (set_local $h0 (i64.mul (get_local $f0) (get_local $sn)))
+    (set_local $h1 (i64.mul (get_local $f1) (get_local $sn)))
+    (set_local $h2 (i64.mul (get_local $f2) (get_local $sn)))
+    (set_local $h3 (i64.mul (get_local $f3) (get_local $sn)))
+    (set_local $h4 (i64.mul (get_local $f4) (get_local $sn)))
+    (set_local $h5 (i64.mul (get_local $f5) (get_local $sn)))
+    (set_local $h6 (i64.mul (get_local $f6) (get_local $sn)))
+    (set_local $h7 (i64.mul (get_local $f7) (get_local $sn)))
+    (set_local $h8 (i64.mul (get_local $f8) (get_local $sn)))
+    (set_local $h9 (i64.mul (get_local $f9) (get_local $sn)))
+
+    (set_local $carry9 (i64.shr_s (i64.add (get_local $h9) (i64.shl (i64.const 1) (i64.const 24))) (i64.const 25)))
+    (set_local $h0 (i64.add (get_local $h0) (i64.mul (get_local $carry9) (i64.const 19))))
+    (set_local $h9 (i64.sub (get_local $h9) (i64.mul (get_local $carry9) (i64.shl (i64.const 1) (i64.const 25)))))
+
+    (set_local $carry1 (i64.shr_s (i64.add (get_local $h1) (i64.shl (i64.const 1) (i64.const 24))) (i64.const 25)))
+    (set_local $h2 (i64.add (get_local $h2) (get_local $carry1)))
+    (set_local $h1 (i64.sub (get_local $h1) (i64.mul (get_local $carry1) (i64.shl (i64.const 1) (i64.const 25)))))
+
+    (set_local $carry3 (i64.shr_s (i64.add (get_local $h3) (i64.shl (i64.const 1) (i64.const 24))) (i64.const 25)))
+    (set_local $h4 (i64.add (get_local $h4) (get_local $carry3)))
+    (set_local $h3 (i64.sub (get_local $h3) (i64.mul (get_local $carry3) (i64.shl (i64.const 1) (i64.const 25)))))
+
+    (set_local $carry5 (i64.shr_s (i64.add (get_local $h5) (i64.shl (i64.const 1) (i64.const 24))) (i64.const 25)))
+    (set_local $h6 (i64.add (get_local $h6) (get_local $carry5)))
+    (set_local $h5 (i64.sub (get_local $h5) (i64.mul (get_local $carry5) (i64.shl (i64.const 1) (i64.const 25)))))
+
+    (set_local $carry7 (i64.shr_s (i64.add (get_local $h7) (i64.shl (i64.const 1) (i64.const 24))) (i64.const 25)))
+    (set_local $h8 (i64.add (get_local $h8) (get_local $carry7)))
+    (set_local $h7 (i64.sub (get_local $h7) (i64.mul (get_local $carry7) (i64.shl (i64.const 1) (i64.const 25)))))
+
+    (set_local $carry0 (i64.shr_s (i64.add (get_local $h0) (i64.shl (i64.const 1) (i64.const 25))) (i64.const 26)))
+    (set_local $h1 (i64.add (get_local $h1) (get_local $carry0)))
+    (set_local $h0 (i64.sub (get_local $h0) (i64.mul (get_local $carry0) (i64.shl (i64.const 1) (i64.const 26)))))
+
+    (set_local $carry2 (i64.shr_s (i64.add (get_local $h2) (i64.shl (i64.const 1) (i64.const 25))) (i64.const 26)))
+    (set_local $h3 (i64.add (get_local $h3) (get_local $carry2)))
+    (set_local $h2 (i64.sub (get_local $h2) (i64.mul (get_local $carry2) (i64.shl (i64.const 1) (i64.const 26)))))
+
+    (set_local $carry4 (i64.shr_s (i64.add (get_local $h4) (i64.shl (i64.const 1) (i64.const 25))) (i64.const 26)))
+    (set_local $h5 (i64.add (get_local $h5) (get_local $carry4)))
+    (set_local $h4 (i64.sub (get_local $h4) (i64.mul (get_local $carry4) (i64.shl (i64.const 1) (i64.const 26)))))
+
+    (set_local $carry6 (i64.shr_s (i64.add (get_local $h6) (i64.shl (i64.const 1) (i64.const 25))) (i64.const 26)))
+    (set_local $h7 (i64.add (get_local $h7) (get_local $carry6)))
+    (set_local $h6 (i64.sub (get_local $h6) (i64.mul (get_local $carry6) (i64.shl (i64.const 1) (i64.const 26)))))
+
+    (set_local $carry8 (i64.shr_s (i64.add (get_local $h8) (i64.shl (i64.const 1) (i64.const 25))) (i64.const 26)))
+    (set_local $h9 (i64.add (get_local $h9) (get_local $carry8)))
+    (set_local $h8 (i64.sub (get_local $h8) (i64.mul (get_local $carry8) (i64.shl (i64.const 1) (i64.const 26)))))
+
+    (i64.store32 offset=0  (get_local $h) (get_local $h0))
+    (i64.store32 offset=4  (get_local $h) (get_local $h1))
+    (i64.store32 offset=8  (get_local $h) (get_local $h2))
+    (i64.store32 offset=12 (get_local $h) (get_local $h3))
+    (i64.store32 offset=16 (get_local $h) (get_local $h4))
+    (i64.store32 offset=20 (get_local $h) (get_local $h5))
+    (i64.store32 offset=24 (get_local $h) (get_local $h6))
+    (i64.store32 offset=28 (get_local $h) (get_local $h7))
+    (i64.store32 offset=32 (get_local $h) (get_local $h8))
+    (i64.store32 offset=36 (get_local $h) (get_local $h9)))
+
+  (func $fe25519_mul32 (export "fe25519_mul32") (param $h i32) (param $f i32) (param $n i32)
+    (i64.load32_u offset=0  (get_local $f))
+    (i64.load32_u offset=4  (get_local $f))
+    (i64.load32_u offset=8  (get_local $f))
+    (i64.load32_u offset=12 (get_local $f))
+    (i64.load32_u offset=16 (get_local $f))
+    (i64.load32_u offset=20 (get_local $f))
+    (i64.load32_u offset=24 (get_local $f))
+    (i64.load32_u offset=28 (get_local $f))
+    (i64.load32_u offset=32 (get_local $f))
+    (i64.load32_u offset=36 (get_local $f))
+    (get_local $n)
+    (get_local $h)
+    (call $fe_mul32)))