diff --git a/lib/smalltalk/test.sh b/lib/smalltalk/test.sh
new file mode 100755
index 00000000..3f6bf531
--- /dev/null
+++ b/lib/smalltalk/test.sh
@@ -0,0 +1,99 @@
+#!/usr/bin/env bash
+# Fast Smalltalk-on-SX test runner — pipes directly to sx_server.exe.
+# Mirrors lib/haskell/test.sh.
+#
+# Usage:
+#   bash lib/smalltalk/test.sh                     # run all tests
+#   bash lib/smalltalk/test.sh -v                  # verbose
+#   bash lib/smalltalk/test.sh tests/tokenize.sx   # run one file
+
+set -uo pipefail
+cd "$(git rev-parse --show-toplevel)"
+
+SX_SERVER="hosts/ocaml/_build/default/bin/sx_server.exe"
+if [ ! -x "$SX_SERVER" ]; then
+  MAIN_ROOT=$(git worktree list | head -1 | awk '{print $1}')
+  if [ -x "$MAIN_ROOT/$SX_SERVER" ]; then
+    SX_SERVER="$MAIN_ROOT/$SX_SERVER"
+  else
+    echo "ERROR: sx_server.exe not found. Run: cd hosts/ocaml && dune build"
+    exit 1
+  fi
+fi
+
+VERBOSE=""
+FILES=()
+for arg in "$@"; do
+  case "$arg" in
+    -v|--verbose) VERBOSE=1 ;;
+    *) FILES+=("$arg") ;;
+  esac
+done
+
+if [ ${#FILES[@]} -eq 0 ]; then
+  mapfile -t FILES < <(find lib/smalltalk/tests -maxdepth 2 -name '*.sx' | sort)
+fi
+
+TOTAL_PASS=0
+TOTAL_FAIL=0
+FAILED_FILES=()
+
+for FILE in "${FILES[@]}"; do
+  [ -f "$FILE" ] || { echo "skip $FILE (not found)"; continue; }
+  TMPFILE=$(mktemp)
+  cat > "$TMPFILE" <<EPOCHS
+(epoch 1)
+(load "lib/smalltalk/tokenizer.sx")
+(epoch 2)
+(load "$FILE")
+(epoch 3)
+(eval "(list st-test-pass st-test-fail)")
+EPOCHS
+
+  OUTPUT=$(timeout 60 "$SX_SERVER" < "$TMPFILE" 2>&1 || true)
+  rm -f "$TMPFILE"
+
+  LINE=$(echo "$OUTPUT" | awk '/^\(ok-len 3 / {getline; print; exit}')
+  if [ -z "$LINE" ]; then
+    LINE=$(echo "$OUTPUT" | grep -E '^\(ok 3 \([0-9]+ [0-9]+\)\)' | tail -1 \
+            | sed -E 's/^\(ok 3 //; s/\)$//')
+  fi
+  if [ -z "$LINE" ]; then
+    echo "X $FILE: could not extract summary"
+    echo "$OUTPUT" | tail -30
+    TOTAL_FAIL=$((TOTAL_FAIL + 1))
+    FAILED_FILES+=("$FILE")
+    continue
+  fi
+  P=$(echo "$LINE" | sed -E 's/^\(([0-9]+) ([0-9]+)\).*/\1/')
+  F=$(echo "$LINE" | sed -E 's/^\(([0-9]+) ([0-9]+)\).*/\2/')
+  TOTAL_PASS=$((TOTAL_PASS + P))
+  TOTAL_FAIL=$((TOTAL_FAIL + F))
+  if [ "$F" -gt 0 ]; then
+    FAILED_FILES+=("$FILE")
+    printf 'X %-40s %d/%d\n' "$FILE" "$P" "$((P+F))"
+    TMPFILE2=$(mktemp)
+    cat > "$TMPFILE2" <<EPOCHS
+(epoch 1)
+(load "lib/smalltalk/tokenizer.sx")
+(epoch 2)
+(load "$FILE")
+(epoch 3)
+(eval "(map (fn (f) (get f :name)) st-test-fails)")
+EPOCHS
+    FAILS=$(timeout 60 "$SX_SERVER" < "$TMPFILE2" 2>&1 | grep -E '^\(ok 3 ' || true)
+    rm -f "$TMPFILE2"
+    echo "    $FAILS"
+  elif [ "$VERBOSE" = "1" ]; then
+    printf 'OK %-40s %d passed\n' "$FILE" "$P"
+  fi
+done
+
+TOTAL=$((TOTAL_PASS + TOTAL_FAIL))
+if [ $TOTAL_FAIL -eq 0 ]; then
+  echo "OK $TOTAL_PASS/$TOTAL smalltalk-on-sx tests passed"
+else
+  echo "FAIL $TOTAL_PASS/$TOTAL passed, $TOTAL_FAIL failed in: ${FAILED_FILES[*]}"
+fi
+
+[ $TOTAL_FAIL -eq 0 ]
diff --git a/lib/smalltalk/tests/tokenize.sx b/lib/smalltalk/tests/tokenize.sx
new file mode 100644
index 00000000..23f5fdb3
--- /dev/null
+++ b/lib/smalltalk/tests/tokenize.sx
@@ -0,0 +1,362 @@
+;; Smalltalk tokenizer tests.
+;;
+;; Lightweight runner: each test checks actual vs expected with structural
+;; equality and accumulates pass/fail counters. Final summary read by
+;; lib/smalltalk/test.sh.
+
+(define
+  st-deep=?
+  (fn
+    (a b)
+    (cond
+      ((= a b) true)
+      ((and (dict? a) (dict? b))
+        (let
+          ((ak (keys a)) (bk (keys b)))
+          (if
+            (not (= (len ak) (len bk)))
+            false
+            (every?
+              (fn
+                (k)
+                (and (has-key? b k) (st-deep=? (get a k) (get b k))))
+              ak))))
+      ((and (list? a) (list? b))
+        (if
+          (not (= (len a) (len b)))
+          false
+          (let
+            ((i 0) (ok true))
+            (begin
+              (define
+                de-loop
+                (fn
+                  ()
+                  (when
+                    (and ok (< i (len a)))
+                    (begin
+                      (when
+                        (not (st-deep=? (nth a i) (nth b i)))
+                        (set! ok false))
+                      (set! i (+ i 1))
+                      (de-loop)))))
+              (de-loop)
+              ok))))
+      (:else false))))
+
+(define st-test-pass 0)
+(define st-test-fail 0)
+(define st-test-fails (list))
+
+(define
+  st-test
+  (fn
+    (name actual expected)
+    (if
+      (st-deep=? actual expected)
+      (set! st-test-pass (+ st-test-pass 1))
+      (begin
+        (set! st-test-fail (+ st-test-fail 1))
+        (append! st-test-fails {:actual actual :expected expected :name name})))))
+
+;; Strip eof and project to just :type/:value.
+(define
+  st-toks
+  (fn
+    (src)
+    (map
+      (fn (tok) {:type (get tok :type) :value (get tok :value)})
+      (filter
+        (fn (tok) (not (= (get tok :type) "eof")))
+        (st-tokenize src)))))
+
+;; ── 1. Whitespace / empty ──
+(st-test "empty input" (st-toks "") (list))
+(st-test "all whitespace" (st-toks "   \t\n  ") (list))
+
+;; ── 2. Identifiers ──
+(st-test
+  "lowercase ident"
+  (st-toks "foo")
+  (list {:type "ident" :value "foo"}))
+
+(st-test
+  "capitalised ident"
+  (st-toks "Foo")
+  (list {:type "ident" :value "Foo"}))
+
+(st-test
+  "underscore ident"
+  (st-toks "_x")
+  (list {:type "ident" :value "_x"}))
+
+(st-test
+  "digits in ident"
+  (st-toks "foo123")
+  (list {:type "ident" :value "foo123"}))
+
+(st-test
+  "two idents separated"
+  (st-toks "foo bar")
+  (list {:type "ident" :value "foo"} {:type "ident" :value "bar"}))
+
+;; ── 3. Keyword selectors ──
+(st-test
+  "keyword selector"
+  (st-toks "foo:")
+  (list {:type "keyword" :value "foo:"}))
+
+(st-test
+  "keyword call"
+  (st-toks "x at: 1")
+  (list
+    {:type "ident" :value "x"}
+    {:type "keyword" :value "at:"}
+    {:type "number" :value 1}))
+
+(st-test
+  "two-keyword chain stays separate"
+  (st-toks "at: 1 put: 2")
+  (list
+    {:type "keyword" :value "at:"}
+    {:type "number" :value 1}
+    {:type "keyword" :value "put:"}
+    {:type "number" :value 2}))
+
+(st-test
+  "ident then assign — not a keyword"
+  (st-toks "x := 1")
+  (list
+    {:type "ident" :value "x"}
+    {:type "assign" :value ":="}
+    {:type "number" :value 1}))
+
+;; ── 4. Numbers ──
+(st-test
+  "integer"
+  (st-toks "42")
+  (list {:type "number" :value 42}))
+
+(st-test
+  "float"
+  (st-toks "3.14")
+  (list {:type "number" :value 3.14}))
+
+(st-test
+  "hex radix"
+  (st-toks "16rFF")
+  (list
+    {:type "number"
+     :value
+     {:radix 16 :digits "FF" :value 255 :kind "radix"}}))
+
+(st-test
+  "binary radix"
+  (st-toks "2r1011")
+  (list
+    {:type "number"
+     :value
+     {:radix 2 :digits "1011" :value 11 :kind "radix"}}))
+
+(st-test
+  "exponent"
+  (st-toks "1e3")
+  (list {:type "number" :value 1000}))
+
+(st-test
+  "negative exponent (parser handles minus)"
+  (st-toks "1.5e-2")
+  (list {:type "number" :value 0.015}))
+
+;; ── 5. Strings ──
+(st-test
+  "simple string"
+  (st-toks "'hi'")
+  (list {:type "string" :value "hi"}))
+
+(st-test
+  "empty string"
+  (st-toks "''")
+  (list {:type "string" :value ""}))
+
+(st-test
+  "doubled-quote escape"
+  (st-toks "'a''b'")
+  (list {:type "string" :value "a'b"}))
+
+;; ── 6. Characters ──
+(st-test
+  "char literal letter"
+  (st-toks "$a")
+  (list {:type "char" :value "a"}))
+
+(st-test
+  "char literal punct"
+  (st-toks "$$")
+  (list {:type "char" :value "$"}))
+
+(st-test
+  "char literal space"
+  (st-toks "$ ")
+  (list {:type "char" :value " "}))
+
+;; ── 7. Symbols ──
+(st-test
+  "symbol ident"
+  (st-toks "#foo")
+  (list {:type "symbol" :value "foo"}))
+
+(st-test
+  "symbol binary"
+  (st-toks "#+")
+  (list {:type "symbol" :value "+"}))
+
+(st-test
+  "symbol arrow"
+  (st-toks "#->")
+  (list {:type "symbol" :value "->"}))
+
+(st-test
+  "symbol keyword chain"
+  (st-toks "#at:put:")
+  (list {:type "symbol" :value "at:put:"}))
+
+(st-test
+  "quoted symbol with spaces"
+  (st-toks "#'foo bar'")
+  (list {:type "symbol" :value "foo bar"}))
+
+;; ── 8. Literal arrays / byte arrays ──
+(st-test
+  "literal array open"
+  (st-toks "#(1 2)")
+  (list
+    {:type "array-open" :value "#("}
+    {:type "number" :value 1}
+    {:type "number" :value 2}
+    {:type "rparen" :value ")"}))
+
+(st-test
+  "byte array open"
+  (st-toks "#[1 2 3]")
+  (list
+    {:type "byte-array-open" :value "#["}
+    {:type "number" :value 1}
+    {:type "number" :value 2}
+    {:type "number" :value 3}
+    {:type "rbracket" :value "]"}))
+
+;; ── 9. Binary selectors ──
+(st-test "plus" (st-toks "+") (list {:type "binary" :value "+"}))
+(st-test "minus" (st-toks "-") (list {:type "binary" :value "-"}))
+(st-test "star" (st-toks "*") (list {:type "binary" :value "*"}))
+(st-test "double-equal" (st-toks "==") (list {:type "binary" :value "=="}))
+(st-test "leq" (st-toks "<=") (list {:type "binary" :value "<="}))
+(st-test "geq" (st-toks ">=") (list {:type "binary" :value ">="}))
+(st-test "neq" (st-toks "~=") (list {:type "binary" :value "~="}))
+(st-test "arrow" (st-toks "->") (list {:type "binary" :value "->"}))
+(st-test "comma" (st-toks ",") (list {:type "binary" :value ","}))
+
+(st-test
+  "binary in expression"
+  (st-toks "a + b")
+  (list
+    {:type "ident" :value "a"}
+    {:type "binary" :value "+"}
+    {:type "ident" :value "b"}))
+
+;; ── 10. Punctuation ──
+(st-test "lparen" (st-toks "(") (list {:type "lparen" :value "("}))
+(st-test "rparen" (st-toks ")") (list {:type "rparen" :value ")"}))
+(st-test "lbracket" (st-toks "[") (list {:type "lbracket" :value "["}))
+(st-test "rbracket" (st-toks "]") (list {:type "rbracket" :value "]"}))
+(st-test "lbrace" (st-toks "{") (list {:type "lbrace" :value "{"}))
+(st-test "rbrace" (st-toks "}") (list {:type "rbrace" :value "}"}))
+(st-test "period" (st-toks ".") (list {:type "period" :value "."}))
+(st-test "semi" (st-toks ";") (list {:type "semi" :value ";"}))
+(st-test "bar" (st-toks "|") (list {:type "bar" :value "|"}))
+(st-test "caret" (st-toks "^") (list {:type "caret" :value "^"}))
+(st-test "bang" (st-toks "!") (list {:type "bang" :value "!"}))
+(st-test "colon" (st-toks ":") (list {:type "colon" :value ":"}))
+(st-test "assign" (st-toks ":=") (list {:type "assign" :value ":="}))
+
+;; ── 11. Comments ──
+(st-test "comment skipped" (st-toks "\"hello\"") (list))
+(st-test
+  "comment between tokens"
+  (st-toks "a \"comment\" b")
+  (list {:type "ident" :value "a"} {:type "ident" :value "b"}))
+(st-test
+  "multi-line comment"
+  (st-toks "\"line1\nline2\"42")
+  (list {:type "number" :value 42}))
+
+;; ── 12. Compound expressions ──
+(st-test
+  "block with params"
+  (st-toks "[:a :b | a + b]")
+  (list
+    {:type "lbracket" :value "["}
+    {:type "colon" :value ":"}
+    {:type "ident" :value "a"}
+    {:type "colon" :value ":"}
+    {:type "ident" :value "b"}
+    {:type "bar" :value "|"}
+    {:type "ident" :value "a"}
+    {:type "binary" :value "+"}
+    {:type "ident" :value "b"}
+    {:type "rbracket" :value "]"}))
+
+(st-test
+  "cascade"
+  (st-toks "x m1; m2")
+  (list
+    {:type "ident" :value "x"}
+    {:type "ident" :value "m1"}
+    {:type "semi" :value ";"}
+    {:type "ident" :value "m2"}))
+
+(st-test
+  "method body return"
+  (st-toks "^ self foo")
+  (list
+    {:type "caret" :value "^"}
+    {:type "ident" :value "self"}
+    {:type "ident" :value "foo"}))
+
+(st-test
+  "class declaration head"
+  (st-toks "Object subclass: #Foo")
+  (list
+    {:type "ident" :value "Object"}
+    {:type "keyword" :value "subclass:"}
+    {:type "symbol" :value "Foo"}))
+
+(st-test
+  "temp declaration"
+  (st-toks "| t1 t2 |")
+  (list
+    {:type "bar" :value "|"}
+    {:type "ident" :value "t1"}
+    {:type "ident" :value "t2"}
+    {:type "bar" :value "|"}))
+
+(st-test
+  "chunk separator"
+  (st-toks "Foo bar !")
+  (list
+    {:type "ident" :value "Foo"}
+    {:type "ident" :value "bar"}
+    {:type "bang" :value "!"}))
+
+(st-test
+  "keyword call with binary precedence"
+  (st-toks "x foo: 1 + 2")
+  (list
+    {:type "ident" :value "x"}
+    {:type "keyword" :value "foo:"}
+    {:type "number" :value 1}
+    {:type "binary" :value "+"}
+    {:type "number" :value 2}))
+
+(list st-test-pass st-test-fail)
diff --git a/lib/smalltalk/tokenizer.sx b/lib/smalltalk/tokenizer.sx
new file mode 100644
index 00000000..e2e47a50
--- /dev/null
+++ b/lib/smalltalk/tokenizer.sx
@@ -0,0 +1,366 @@
+;; Smalltalk tokenizer.
+;;
+;; Token types:
+;;   ident      identifier (foo, Foo, _x)
+;;   keyword    selector keyword (foo:) — value is "foo:" with the colon
+;;   binary     binary selector chars run together (+, ==, ->, <=, ~=, ...)
+;;   number     integer or float; radix integers like 16rFF supported
+;;   string     'hello''world' style
+;;   char       $c
+;;   symbol     #foo, #foo:bar:, #+, #'with spaces'
+;;   array-open  #(
+;;   byte-array-open  #[
+;;   lparen rparen lbracket rbracket lbrace rbrace
+;;   period semi bar caret colon assign bang
+;;   eof
+;;
+;; Comments "…" are skipped.
+
+(define st-make-token (fn (type value pos) {:type type :value value :pos pos}))
+
+(define st-digit? (fn (c) (and (not (= c nil)) (>= c "0") (<= c "9"))))
+
+(define
+  st-letter?
+  (fn
+    (c)
+    (and
+      (not (= c nil))
+      (or (and (>= c "a") (<= c "z")) (and (>= c "A") (<= c "Z"))))))
+
+(define st-ident-start? (fn (c) (or (st-letter? c) (= c "_"))))
+
+(define st-ident-char? (fn (c) (or (st-ident-start? c) (st-digit? c))))
+
+(define st-ws? (fn (c) (or (= c " ") (= c "\t") (= c "\n") (= c "\r"))))
+
+(define
+  st-binary-chars
+  (list "+" "-" "*" "/" "\\" "~" "<" ">" "=" "@" "%" "&" "?" ","))
+
+(define
+  st-binary-char?
+  (fn (c) (and (not (= c nil)) (contains? st-binary-chars c))))
+
+(define
+  st-radix-digit?
+  (fn
+    (c)
+    (and
+      (not (= c nil))
+      (or (st-digit? c) (and (>= c "A") (<= c "Z"))))))
+
+(define
+  st-tokenize
+  (fn
+    (src)
+    (let
+      ((tokens (list)) (pos 0) (src-len (len src)))
+      (define
+        pk
+        (fn
+          (offset)
+          (if (< (+ pos offset) src-len) (nth src (+ pos offset)) nil)))
+      (define cur (fn () (pk 0)))
+      (define advance! (fn (n) (set! pos (+ pos n))))
+      (define
+        push!
+        (fn
+          (type value start)
+          (append! tokens (st-make-token type value start))))
+      (define
+        skip-comment!
+        (fn
+          ()
+          (cond
+            ((>= pos src-len) nil)
+            ((= (cur) "\"") (advance! 1))
+            (else (begin (advance! 1) (skip-comment!))))))
+      (define
+        skip-ws!
+        (fn
+          ()
+          (cond
+            ((>= pos src-len) nil)
+            ((st-ws? (cur)) (begin (advance! 1) (skip-ws!)))
+            ((= (cur) "\"") (begin (advance! 1) (skip-comment!) (skip-ws!)))
+            (else nil))))
+      (define
+        read-ident-chars!
+        (fn
+          ()
+          (when
+            (and (< pos src-len) (st-ident-char? (cur)))
+            (begin (advance! 1) (read-ident-chars!)))))
+      (define
+        read-decimal-digits!
+        (fn
+          ()
+          (when
+            (and (< pos src-len) (st-digit? (cur)))
+            (begin (advance! 1) (read-decimal-digits!)))))
+      (define
+        read-radix-digits!
+        (fn
+          ()
+          (when
+            (and (< pos src-len) (st-radix-digit? (cur)))
+            (begin (advance! 1) (read-radix-digits!)))))
+      (define
+        read-exp-part!
+        (fn
+          ()
+          (when
+            (and
+              (< pos src-len)
+              (or (= (cur) "e") (= (cur) "E"))
+              (let
+                ((p1 (pk 1)) (p2 (pk 2)))
+                (or
+                  (st-digit? p1)
+                  (and (or (= p1 "+") (= p1 "-")) (st-digit? p2)))))
+            (begin
+              (advance! 1)
+              (when
+                (and (< pos src-len) (or (= (cur) "+") (= (cur) "-")))
+                (advance! 1))
+              (read-decimal-digits!)))))
+      (define
+        read-number
+        (fn
+          (start)
+          (begin
+            (read-decimal-digits!)
+            (cond
+              ((and (< pos src-len) (= (cur) "r"))
+                (let
+                  ((base-str (slice src start pos)))
+                  (begin
+                    (advance! 1)
+                    (let
+                      ((rstart pos))
+                      (begin
+                        (read-radix-digits!)
+                        (let
+                          ((digits (slice src rstart pos)))
+                          {:radix (parse-number base-str)
+                           :digits digits
+                           :value (parse-radix base-str digits)
+                           :kind "radix"}))))))
+              ((and
+                  (< pos src-len)
+                  (= (cur) ".")
+                  (st-digit? (pk 1)))
+                (begin
+                  (advance! 1)
+                  (read-decimal-digits!)
+                  (read-exp-part!)
+                  (parse-number (slice src start pos))))
+              (else
+                (begin
+                  (read-exp-part!)
+                  (parse-number (slice src start pos))))))))
+      (define
+        parse-radix
+        (fn
+          (base-str digits)
+          (let
+            ((base (parse-number base-str))
+              (chars digits)
+              (n-len (len digits))
+              (idx 0)
+              (acc 0))
+            (begin
+              (define
+                rd-loop
+                (fn
+                  ()
+                  (when
+                    (< idx n-len)
+                    (let
+                      ((c (nth chars idx)))
+                      (let
+                        ((d (cond
+                              ((and (>= c "0") (<= c "9")) (- (char-code c) 48))
+                              ((and (>= c "A") (<= c "Z")) (- (char-code c) 55))
+                              (else 0))))
+                        (begin
+                          (set! acc (+ (* acc base) d))
+                          (set! idx (+ idx 1))
+                          (rd-loop)))))))
+              (rd-loop)
+              acc))))
+      (define
+        read-string
+        (fn
+          ()
+          (let
+            ((chars (list)))
+            (begin
+              (advance! 1)
+              (define
+                loop
+                (fn
+                  ()
+                  (cond
+                    ((>= pos src-len) nil)
+                    ((= (cur) "'")
+                      (cond
+                        ((= (pk 1) "'")
+                          (begin
+                            (append! chars "'")
+                            (advance! 2)
+                            (loop)))
+                        (else (advance! 1))))
+                    (else
+                      (begin (append! chars (cur)) (advance! 1) (loop))))))
+              (loop)
+              (join "" chars)))))
+      (define
+        read-binary-run!
+        (fn
+          ()
+          (let
+            ((start pos))
+            (begin
+              (define
+                bin-loop
+                (fn
+                  ()
+                  (when
+                    (and (< pos src-len) (st-binary-char? (cur)))
+                    (begin (advance! 1) (bin-loop)))))
+              (bin-loop)
+              (slice src start pos)))))
+      (define
+        read-symbol
+        (fn
+          (start)
+          (cond
+            ;; Quoted symbol: #'whatever'
+            ((= (cur) "'")
+              (let ((s (read-string))) (push! "symbol" s start)))
+            ;; Binary-char symbol: #+, #==, #->, #|
+            ((or (st-binary-char? (cur)) (= (cur) "|"))
+              (let ((b (read-binary-run!)))
+                (cond
+                  ((= b "")
+                    ;; lone | wasn't binary; consume it
+                    (begin (advance! 1) (push! "symbol" "|" start)))
+                  (else (push! "symbol" b start)))))
+            ;; Identifier or keyword chain: #foo, #foo:bar:
+            ((st-ident-start? (cur))
+              (let ((id-start pos))
+                (begin
+                  (read-ident-chars!)
+                  (define
+                    kw-loop
+                    (fn
+                      ()
+                      (when
+                        (and (< pos src-len) (= (cur) ":"))
+                        (begin
+                          (advance! 1)
+                          (when
+                            (and (< pos src-len) (st-ident-start? (cur)))
+                            (begin (read-ident-chars!) (kw-loop)))))))
+                  (kw-loop)
+                  (push! "symbol" (slice src id-start pos) start))))
+            (else
+              (error
+                (str "st-tokenize: bad symbol at " pos))))))
+      (define
+        step
+        (fn
+          ()
+          (begin
+            (skip-ws!)
+            (when
+              (< pos src-len)
+              (let
+                ((start pos) (c (cur)))
+                (cond
+                  ;; Identifier or keyword
+                  ((st-ident-start? c)
+                    (begin
+                      (read-ident-chars!)
+                      (let
+                        ((word (slice src start pos)))
+                        (cond
+                          ;; ident immediately followed by ':' (and not ':=') => keyword
+                          ((and
+                              (< pos src-len)
+                              (= (cur) ":")
+                              (not (= (pk 1) "=")))
+                            (begin
+                              (advance! 1)
+                              (push!
+                                "keyword"
+                                (str word ":")
+                                start)))
+                          (else (push! "ident" word start))))
+                      (step)))
+                  ;; Number
+                  ((st-digit? c)
+                    (let
+                      ((v (read-number start)))
+                      (begin (push! "number" v start) (step))))
+                  ;; String
+                  ((= c "'")
+                    (let
+                      ((s (read-string)))
+                      (begin (push! "string" s start) (step))))
+                  ;; Character literal
+                  ((= c "$")
+                    (cond
+                      ((>= (+ pos 1) src-len)
+                        (error (str "st-tokenize: $ at end of input")))
+                      (else
+                        (begin
+                          (advance! 1)
+                          (push! "char" (cur) start)
+                          (advance! 1)
+                          (step)))))
+                  ;; Symbol or array literal
+                  ((= c "#")
+                    (cond
+                      ((= (pk 1) "(")
+                        (begin (advance! 2) (push! "array-open" "#(" start) (step)))
+                      ((= (pk 1) "[")
+                        (begin (advance! 2) (push! "byte-array-open" "#[" start) (step)))
+                      (else
+                        (begin (advance! 1) (read-symbol start) (step)))))
+                  ;; Assignment :=  or  bare colon
+                  ((= c ":")
+                    (cond
+                      ((= (pk 1) "=")
+                        (begin (advance! 2) (push! "assign" ":=" start) (step)))
+                      (else
+                        (begin (advance! 1) (push! "colon" ":" start) (step)))))
+                  ;; Single-char structural punctuation
+                  ((= c "(") (begin (advance! 1) (push! "lparen" "(" start) (step)))
+                  ((= c ")") (begin (advance! 1) (push! "rparen" ")" start) (step)))
+                  ((= c "[") (begin (advance! 1) (push! "lbracket" "[" start) (step)))
+                  ((= c "]") (begin (advance! 1) (push! "rbracket" "]" start) (step)))
+                  ((= c "{") (begin (advance! 1) (push! "lbrace" "{" start) (step)))
+                  ((= c "}") (begin (advance! 1) (push! "rbrace" "}" start) (step)))
+                  ((= c ".") (begin (advance! 1) (push! "period" "." start) (step)))
+                  ((= c ";") (begin (advance! 1) (push! "semi" ";" start) (step)))
+                  ((= c "|") (begin (advance! 1) (push! "bar" "|" start) (step)))
+                  ((= c "^") (begin (advance! 1) (push! "caret" "^" start) (step)))
+                  ((= c "!") (begin (advance! 1) (push! "bang" "!" start) (step)))
+                  ;; Binary selector run
+                  ((st-binary-char? c)
+                    (let
+                      ((b (read-binary-run!)))
+                      (begin (push! "binary" b start) (step))))
+                  (else
+                    (error
+                      (str
+                        "st-tokenize: unexpected char "
+                        c
+                        " at "
+                        pos)))))))))
+      (step)
+      (push! "eof" nil pos)
+      tokens)))
diff --git a/plans/smalltalk-on-sx.md b/plans/smalltalk-on-sx.md
index 2d4f47f1..51412d67 100644
--- a/plans/smalltalk-on-sx.md
+++ b/plans/smalltalk-on-sx.md
@@ -50,7 +50,7 @@ Core mapping:
 ## Roadmap
 
 ### Phase 1 — tokenizer + parser
-- [ ] Tokenizer: identifiers, keywords (`foo:`), binary selectors (`+`, `==`, `,`, `->`, `~=` etc.), numbers (radix `16r1F`, scaled `1.5s2`), strings `'…''…'`, characters `$c`, symbols `#foo` `#'foo bar'` `#+`, byte arrays `#[1 2 3]`, literal arrays `#(1 #foo 'x')`, comments `"…"`
+- [x] Tokenizer: identifiers, keywords (`foo:`), binary selectors (`+`, `==`, `,`, `->`, `~=` etc.), numbers (radix `16r1F`; **scaled `1.5s2` deferred**), strings `'…''…'`, characters `$c`, symbols `#foo` `#'foo bar'` `#+`, byte arrays `#[1 2 3]` (open token), literal arrays `#(1 #foo 'x')` (open token), comments `"…"`
 - [ ] Parser: chunk format (`! !` separators), class definitions (`Object subclass: #X instanceVariableNames: '…' classVariableNames: '…' …`), method definitions (`extend: #Foo with: 'bar ^self'`), pragmas `<primitive: 1>`, blocks `[:a :b | | t1 t2 | …]`, cascades, message precedence (unary > binary > keyword)
 - [ ] Unit tests in `lib/smalltalk/tests/parse.sx`
 
@@ -107,7 +107,7 @@ Core mapping:
 
 _Newest first. Agent appends on every commit._
 
-- _(none yet)_
+- 2026-04-25: tokenizer + 63 tests (`lib/smalltalk/tokenizer.sx`, `lib/smalltalk/tests/tokenize.sx`, `lib/smalltalk/test.sh`). All token types covered except scaled decimals `1.5s2` (deferred). `#(` and `#[` emit open tokens; literal-array contents lexed as ordinary tokens for the parser to interpret.
 
 ## Blockers