diff --git a/2023/day19.nim b/2023/day19.nim
new file mode 100644
index 0000000..b639b60
--- /dev/null
+++ b/2023/day19.nim
@@ -0,0 +1,150 @@
+import std/algorithm
+import std/bitops
+import std/deques
+import std/enumerate
+import std/heapqueue
+import std/math
+import std/options
+import std/sequtils
+import std/sets
+import std/strscans
+import std/strutils
+import std/sugar
+import std/tables
+import std/terminal
+import std/times
+
+proc dbg[T](v: T): T = (echo(v); v)
+
+func empty[T](v: seq[T]): bool = v.len == 0
+
+proc split[T](v: seq[T], pred: proc(elt: T): bool {.noSideEffect.}): seq[seq[T]] =
+  var current: seq[T]
+  for elt in v:
+    if pred(elt):
+      result.add current
+      current = @[]
+    else:
+      current.add elt
+  result.add current
+
+func join[T](parts: seq[seq[T]], sep: T): seq[T] =
+  if parts.len > 0:
+    result &= parts[0]
+
+  for part in parts[1..^1]:
+    result.add sep
+    result &= part
+
+###################
+
+type Op = enum
+  None
+  Lt
+  Gt
+
+func parseOp(c: char): Op =
+  case c:
+    of '<': Lt
+    of '>': Gt
+    else: raise newException(ValueError, $c)
+
+type Step = tuple
+  attr: char
+  op: Op
+  val: int
+  dest: string
+
+func parseStep(text: string): Step =
+  var op: char
+  if scanf(text, "$c$c$i:$+$.", result.attr, op, result.val, result.dest):
+    result.op = op.parseOp
+    return
+
+  if scanf(text, "$w$.", result.dest):
+    result.op = None
+    return
+
+  raise newException(ValueError, text)
+
+type Workflow = tuple
+  name: string
+  steps: seq[Step]
+
+func parseWorkflow(text: string): Workflow =
+  var steps: string
+  if not scanf(text, "$+{$+}$.", result.name, steps):
+    raise newException(ValueError, text)
+  result.steps = steps.split(',').map(parseStep)
+
+type Part = tuple
+  attrs: Table[char, int]
+
+func parseAttr(text: string): (char, int) =
+  if not scanf(text, "$c=$i$.", result[0], result[1]):
+    raise newException(ValueError, text)
+
+func parseAttrs(text: string): Table[char, int] =
+  for (a, v) in text.split(',').map(parseAttr):
+    result[a] = v
+
+func parsePart(line: string): Part =
+  var attrs: string
+  if not scanf(line, "{$+}$.", attrs):
+    raise newException(ValueError, line)
+  result.attrs = attrs.parseAttrs
+
+type System = tuple
+  workflows: Table[string, Workflow]
+  parts: seq[Part]
+
+func parseSystem(text: string): System =
+  let sections = text.split("\n\n")
+  assert sections.len == 2
+
+  for w in sections[0].strip.splitLines.map(parseWorkflow):
+    result.workflows[w.name] = w
+
+  result.parts = sections[1].strip.splitLines.map(parsePart)
+
+func apply(w: Workflow, p: Part): string =
+  for s in w.steps:
+    case s.op:
+      of None:
+        return s.dest
+      of Lt:
+        if p.attrs[s.attr] < s.val:
+          return s.dest
+      of Gt:
+        if p.attrs[s.attr] > s.val:
+          return s.dest
+
+func accepted(sys: System, part: Part): bool =
+  var w = "in"
+  while w != "R" and w != "A":
+    w = sys.workflows[w].apply(part)
+  w == "A"
+
+func rating(p: Part): int =
+  for v in p.attrs.values:
+    result.inc v
+
+proc part1(input: string): int =
+  let text = readFile(input)
+  let system = parseSystem(text)
+
+  for part in system.parts:
+    if system.accepted(part):
+      result.inc part.rating
+
+###################
+
+proc timed[T](f: () -> T): T =
+  let start = cpuTime()
+  result = f()
+  echo "Time :", cpuTime() - start
+
+echo timed(proc(): int = part1("input/test.txt"))
+echo timed(proc(): int = part1("input/day19.txt"))
+# echo timed(proc(): int = part2("input/test.txt"))
+# echo timed(proc(): int = part2("input/day19.txt"))