Pk2 Extractor Apr 2026

# Decompress if needed (zlib) if flags & 1: data = zlib.decompress(data)

print("Done!") if == " main ": import sys if len(sys.argv) < 3: print("Usage: python pk2_extractor.py <file.pk2> <output_folder>") else: extract_pk2(sys.argv[1], sys.argv[2]) Step 4: Running the Extractor Open a terminal and run:

# Write to disk with open(out_path, "wb") as out_f: out_f.write(data) print(f"Extracted: file_path")

# Save current position to read file name current_pos = f.tell() f.seek(name_offset) file_path = f.read(256).split(b"\x00")[0].decode("utf-8", errors="ignore") f.seek(current_pos) pk2 extractor

– [Your Name]

# Read file data f.seek(file_offset) data = f.read(compressed_size)

| Offset | Size (bytes) | Description | |--------|--------------|-------------| | 0 | 4 | Magic header ( PK20 or PK2 ) | | 4 | 4 | Version (usually 2) | | 8 | 4 | Number of files | | 12 | 4 | Offset to file index table | | 16 | 4 | Unknown/Reserved | | 20 | ... | File index entries | # Decompress if needed (zlib) if flags & 1: data = zlib

import os import struct import zlib def extract_pk2(pk2_path, output_dir): with open(pk2_path, "rb") as f: # Read header magic = f.read(4) if magic not in (b"PK20", b"PK2\x00"): raise ValueError("Not a valid PK2 file")

But here’s the problem: modern Windows doesn’t open PK2 files. Double-clicking does nothing. So what do you do when you need to extract that one weapon texture or edit a quest script?

# Prepare output path out_path = os.path.join(output_dir, file_path) os.makedirs(os.path.dirname(out_path), exist_ok=True) So what do you do when you need

version, num_files, index_offset = struct.unpack("<III", f.read(12)) print(f"Version: version, Files: num_files, Index at: index_offset")

In this post, I’ll walk through the PK2 format, write a lightweight Python extractor from scratch, and show you how to unpack those archives in seconds. After reversing a few sample PK2 files (and thanks to open-source community notes), the format breaks down like this: