diff --git a/src/dat/nbt.c b/src/dat/nbt.c
new file mode 100644
index 0000000..2a6195a
--- /dev/null
+++ b/src/dat/nbt.c
@@ -0,0 +1,32 @@
+#include "nbt.h"
+
+#include <endian.h>
+#include <stddef.h>
+
+#include "../util/types.h"
+
+/* the data structure that functions passes between functions, so it can communicate with the central function */
+struct nbt_procdat {
+	u32 ncomp; // the number of compound tags we've encountered
+};
+
+/* returns the string length from a specific location in the buffer */
+static inline u16 nbt_strlen(u8 const *restrict buf) {
+	return be16toh(*(u16 *)(buf));
+}
+
+int nbt_proc(void **restrict datout, u8 const *restrict buf, size_t len) {
+	struct nbt_procdat procdat = {0};
+
+	// first byte should be a compound tag
+	if (*buf != NBT_COMPOUND) return 1;
+
+	// ignore the first tag + its name, so we start with the juicy data
+	uint tmp = nbt_strlen(buf + 1) + 3;
+	buf += tmp;
+	len -= tmp;
+	procdat.ncomp++;
+
+	// TODO: finish function
+	return 0;
+}
diff --git a/src/dat/nbt.h b/src/dat/nbt.h
index 31a2677..1fc6c5c 100644
--- a/src/dat/nbt.h
+++ b/src/dat/nbt.h
@@ -2,6 +2,11 @@
 // Licensed under the MIT Licence. See LICENSE for details
 #pragma once
 
+#include <stdbool.h>
+#include <stdlib.h>
+
+#include "../util/types.h"
+
 /* NBT (named binary tag) is a tree data structure. Tags have a numeric type ID, name and a payload.
  * NBT files are a compressed `compound` tag. GZip is the compression used in most cases,
  * in some (rare) cases it's stored uncompressed.
@@ -27,3 +32,5 @@ enum nbt_tagid {
 	NBT_ARR_I32 = 0x0B,  // starts with a i32, denoting size, followed by the i32 data
 	NBT_ARR_I64 = 0x0C,  // starts with a i32, denoting size, followed by the u32 data
 };
+
+int nbt_proc(void **restrict datout, u8 const *restrict buf, size_t len);