Download Quixel Megascans with the ability to dictate what resolution and file type you want to download. Also checks for corrupt zip files. Tested with Surfaces and 3d Assets.

download_megascans.py

import zipfile
import os
import json
import concurrent.futures
import requests
import shutil
import time

######## INITIAL SETUP ########
# Define the token, download path, and target category
TOKEN = "Your Token Here"

# Update with the correct path to the directory
# Example: "C:/Users/MyAccount/Documents/QuixelZips"
DOWNLOAD_PATH = ""

# Update with the correct path to the ms_asset_categories.json file, not the
# directory.
# Example: C:/Users/MyAccount/Documents/QuixelZips/ms_asset_categoies.json"
JSON_FILE_PATH = "./ms_asset_categories.json"  

# Pick a folder to store the download cache at (a text file). Then create the
# text file cache.txt there. Paste it below.
# IE: "C:/Users/MyAccount/Documents/Quixel/cache.txt"
CACHE_FILE_PATH = "./cache.txt"

# Download from https://github.com/WAUthethird/quixel-megascans-scripts -->
# "complete_asset_metadata.tar.zst", then unzip it.
ASSET_METADATA_PATH = "./asset_metadata.json"

## Set target download category/categories.
#working: 3d asset, 3d plant, surface, brush, displacement, imperfection, decal
TARGET_CATEGORY = "3d assets"

# Create subdirectories based on the category
SAVE_IN_SUBDIRECTORIES = True

# Download HighPoly?
HIGHPOLY = False

# Donwload ztool (ZBrush) file?
ZTOOL = False

# Use to overwrite existing cached items. (Example if you want to downlaod a
# different size texture. Or if they messed up and you had to adjust script to
# try again.)
OVERWRITE = False

# Set a limit for the number of items you want to download. 0 means no limit
MAX_ITEM_COUNT = 0

# Specify the image type and resolution for all textures, or for default.
# If a value is not specified, then it will pull from the default item.
# If the resolution specified does not exist for that item (surface, asset, etc)
# then it will default to the highest resolution available for that item.
# So if you specify "8192x8192" but the highest that exists is 2048x2048, then it
# will download 2048x2048 instead. Note that some textures are not square, and
# so the highest resolution available will still download. IE could be 2048x1024

MIME_TEXTURE_TYPES = {
    "default":          {"mimeType": "image/jpeg", "resolution": "4096x4096"},
    # "albedo":           {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "ao":               {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "brush":            {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "bump":             {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "cavity":           {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "curvature":        {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "diffuse":          {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    "displacement":     {"mimeType": "image/x-exr", "resolution": "8192x8192"},
    # "f":                {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "fuzz":             {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "gloss":            {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "mask":             {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "metalness":        {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    "normal":           {"mimeType": "image/x-exr", "resolution": "4096x4096"},
    # "normalbump":       {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "normalobject":     {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "occlusion":        {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "opacity":          {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "roughness":        {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "specular":         {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "thickness":        {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "translucency":     {"mimeType": "image/jpeg", "resolution": "2048x2048"},
    # "transmission":     {"mimeType": "image/jpeg", "resolution": "2048x2048"},
}

######## SETUP - FINISHED ########


# Function to normalize category strings
def normalize_category(category):
    return category.strip().lower().rstrip('s')

# Function to convert categories to a space-separated string
def categories_to_string(categories, separator=" "):
    result = []
    if isinstance(categories, dict):
        for key, value in categories.items():
            if isinstance(value, dict) and value:
                subcategories = categories_to_string(value, separator)
                if subcategories:
                    result.append(f"{key}{separator}{subcategories}")
                else:
                    result.append(key)
            elif isinstance(value, list):
                for item in value:
                    result.append(normalize_category(item))
            else:
                result.append(normalize_category(key))
    return separator.join(result)

    
def get_asset_download_id(asset_id): 
    url = "https://quixel.com/v1/downloads"
    paramToPass = get_asset_payload(asset_id)

    trys = 0
    paramLen = len(paramToPass["components"])
    while(trys < paramLen-1):
        trys += 1
        download_url_response = requests.post(
            url,
            headers={"Authorization": "Bearer " + TOKEN},
            json=paramToPass
        )

        if download_url_response.status_code == 200:
            print(f"Found Download URL for {asset_id}")
            return download_url_response.json()["id"]
        elif download_url_response.status_code == 401:
            print((
                "Possible expired token. Please get a new one from "
                "https://quixel.com/megascans/home/ and then update the script.\n"
                "If you just ran the script and downloaded stuff prior to this, "
                "just re-run the script and try again."
            ))
            os._exit(0)
        else:
            print(f"Failed to get asset download url for id: {asset_id}")
            response = download_url_response.json()

            if response["code"] == "INVALID_PAYLOAD":
                if "type not found" in response["msg"]:
                    payload = json.loads(response["msg"].replace(" type not found", ""))
                    if payload:
                        for idx, v in enumerate(paramToPass["components"]):
                            if paramToPass["components"][idx]["type"] == payload["type"]:
                                paramToPass["components"].pop(idx)
                                print("Removed "+payload["type"]+" from payload. Trying again...")
                                break
                else:
                    print("DEBUG_ERROR: " + str(response))
                    return None
            else:
                print("DEBUG_ERROR: " + str(response))
                return None


def check_zip_file(file_path):
    the_zip_file = zipfile.ZipFile(file_path)
    result = the_zip_file.testzip()
    if result is not None:
        print("!! First bad file in zip: %s" % result)

        
def download_asset(download_id, download_directory, asset_id):
    # full_name = asset_metadata["asset_metadata"][download_id]
    os.makedirs(download_directory, exist_ok=True)

    url = f"https://assetdownloads.quixel.com/download/{download_id}?preserveStructure=True&url=https%3A%2F%2Fquixel.com%2Fv1%2Fdownloads"
    
    response = requests.get(url, stream=True)
    
    if response.status_code == 400:
        # Print the response to see what's causing the issue
        print(f"Error 400: {response.text}")

    attempt_count = 0
    delay = 3
    max_attempts = 5

    while attempt_count < max_attempts:
        response = requests.get(url, stream=True)
        if response.status_code == 200:
            content_disposition = response.headers.get("Content-Disposition")
            if content_disposition:
                filename = content_disposition.split("filename=")[-1].strip('"') 
            else:
                filename = download_id
                
            asset = asset_metadata["asset_metadata"].get(asset_id)
            if asset:
                asset_name = asset["name"].strip().replace(" ", "_")
                ext = os.path.splitext(filename)[-1]
                base_path = os.path.dirname(filename)
                filename = os.path.join(base_path, f"{asset_name}_{asset_id}{ext}")
            file_path = os.path.join(download_directory, filename)

            try:
                print(f"Downloading file: {file_path} ###")
                with requests.get(url, stream=True) as r:
                    with open(file_path, 'wb') as f:
                        shutil.copyfileobj(r.raw, f)
                check_zip_file(file_path.replace('\\', '/'))
                return True
            except requests.exceptions.ChunkedEncodingError as e:
                print(f"Error during download: {e}")
                time.sleep(delay)
                delay += 3
                attempt_count += 1
            except zipfile.BadZipFile:
                print((
                    "Bad zip file found, removing zip file and attempting to "
                    "redownload..."
                ))
                os.remove(file_path)
        else:
            print((
                f"ERROR: Failed to download asset {download_id}, status code: "
                f"{response.status_code}\nResponse: {response.text}"
            ))
            return False

    print(f"Exceeded maximum retry attempts for asset {download_id}")
    return False

        
# Function to handle downloading for threading
def download_asset_with_id(asset):
    asset_id = asset["asset_id"]
    path = ""
    if(SAVE_IN_SUBDIRECTORIES):
        path = asset["path"].lower()
    download_id = get_asset_download_id(asset_id)
    if download_id:
        return download_asset(download_id, os.path.join(DOWNLOAD_PATH, path), asset_id)
    else:
        print(f"No download id found for {asset_id}.")
        return False


def get_asset_payload(asset_id):
    asset = [asset for asset in asset_metadata["asset_metadata"].values() if asset["full_metadata"]["id"] == asset_id]
    if len(asset) <=0:
        print(f"Asset {asset_id} not found in asset_metadata.json!")
        return False

    asset = asset[0]

    if "components" in asset["full_metadata"]:
        texture_types = {}
        for component in asset["full_metadata"]["components"]:
            resolutions = []
            for uri in component["uris"]:
                for resolution_dict in uri["resolutions"]:
                    resolution = resolution_dict['resolution']
                    resolutions.append(resolution)
            texture_types[component["type"]] = resolutions
    # Surfaces
    else:
        texture_types = {}
        for map in asset["full_metadata"]["maps"]:
            texture = map["type"]
            if texture not in texture_types:
                texture_types[texture] = []
            texture_types[texture].append(map["resolution"])

    asset_components = []
    for texture, resolutions in texture_types.items():
        resolution_requested = MIME_TEXTURE_TYPES.get(
            texture,
            MIME_TEXTURE_TYPES["default"]
        )["resolution"]
        # Get highest AVAILABLE resolution for that texture
        if resolution_requested not in resolutions:
            highest_value = 0
            for res in resolutions:
                w, h = res.split("x")
                res_sum = int(w) + int(h)
                if res_sum > highest_value:
                    highest_value = res_sum
                    resolution_requested = res
            
        asset_components.append({
            "type": texture,
            "mimeType": MIME_TEXTURE_TYPES.get(texture, MIME_TEXTURE_TYPES["default"])["mimeType"],
            "resolution": resolution_requested,
        })

    payload = {"asset": asset_id,
            "config": {"highpoly": HIGHPOLY,
                       "lowerlod_meshes": True,
                       "lowerlod_normals": True,
                       "ztool": ZTOOL,
                       "brushes": True,
                       "meshMimeType": "application/x-fbx",
                       "albedo_lods": True},
            "components": asset_components}

    return payload


if __name__ == "__main__":
    with open(JSON_FILE_PATH, 'r') as f:
        asset_categories_dict = json.load(f)

    print(f"Trying to match against target category: {TARGET_CATEGORY}")

    # Load cached assets
    cached_assets = set()
    if os.path.exists(CACHE_FILE_PATH):
        with open(CACHE_FILE_PATH, "r") as cache_file:
            cached_assets = set(cache_file.read().splitlines())

    # Normalize target category for matching
    normalized_target_categories = [normalize_category(part) for part in TARGET_CATEGORY.split("/")]

    matching_asset_ids = []
    # Check matches for each asset in the loaded categories
    for asset_id, categories in asset_categories_dict.items():
        # Convert the categories to a single string for matching
        categories_str = categories_to_string(categories)
        categories_path = categories_to_string(categories, "/")
        
        # Check if all parts of target_category exist in the categories string
        matches = all(normalize_category(part) in categories_str.lower() for part in normalized_target_categories)
        
        if matches and asset_id not in cached_assets:
            matching_asset_ids.append({"asset_id": asset_id, "path": categories_path})
        elif matches and asset_id in cached_assets and OVERWRITE:
            matching_asset_ids.append({"asset_id": asset_id, "path": categories_path})

    if not matching_asset_ids:
        print("No new assets found for the target category.")
        exit()
    
    print(f"{len(matching_asset_ids)} assets found.")
    
    # Allow user to decide how many to download at once.
    if MAX_ITEM_COUNT:
        item_count = min(len(matching_asset_ids), MAX_ITEM_COUNT)
    else:
        item_count = len(matching_asset_ids)
    
    confirmation = input((
        f"Do you want to download {item_count} {TARGET_CATEGORY} assets? "
        "(y/n): "
    )).strip().lower()
    
    if confirmation != "y":
        print("Download canceled.")
        exit()

    try:
        with open(ASSET_METADATA_PATH, "r", encoding="utf-8") as f:
            asset_metadata = json.load(f)
    except FileNotFoundError:
        print((
            "Couldn't find asset_metadata.json in the directory you selected, "
            f"{ASSET_METADATA_PATH}. Quitting."
        ))
        exit()

    time_start = time.time()
    with open(CACHE_FILE_PATH, "a+") as cache_file:
        # Use threading for faster downloading
        with concurrent.futures.ThreadPoolExecutor(max_workers=10) as executor:
            download_count = item_count
            futures = {executor.submit(download_asset_with_id, asset): asset for asset in matching_asset_ids[:item_count]}
            for future in concurrent.futures.as_completed(futures):
                asset = futures[future]
                asset_id = asset["asset_id"]
                try:
                    result = future.result()
                    if result:
                        download_count-=1
                        print(f"{download_count} remaining items to download.")
                        # Add the asset to the cache file after successful download
                        cache_file.write(f"{asset_id}\n")
                        cache_file.flush()
                except Exception as e:
                    print(f"Error downloading asset {asset_id}: {e}")
    print(f"Time Taken: {time.time()-time_start} seconds")

get_token.js

// Function to get the value of a specific cookie by name
function getCookie(name) {
    const value = `; ${document.cookie}`;
    const parts = value.split(`; ${name}=`);
    if (parts.length === 2) return parts.pop().split(';').shift();
  }
  
  // Get the auth cookie
  const authCookie = getCookie('auth');
  
  // Parse the auth cookie (it should be a JSON string containing the token)
  if (authCookie) {
    try {
      const authData = JSON.parse(decodeURIComponent(authCookie));
      const authToken = authData.token;
      console.log("Auth Token:", authToken);
    } catch (error) {
      console.error("Error parsing auth cookie:", error);
    }
  } else {
    console.error("Auth cookie not found. Please make sure you are logged in.");
  }

Hi, I think I have followed all the instructions correctly. But I am getting an error that reads,

ModuleNotFoundError: No module named 'requests'

Any idea what i might be doing wrong?

Have you tried pip install requests?

Hey @PolyShifter,

Using the script on Surfaces and going well. Downloading 2k versions and getting about 1000-2000 files per hour (getting faster as the night goes on... probably because of less traffic). I am getting a significant number of 'connection broken' errors though (about 1 out of 10). Any suggestions?

Also, are there any categories that will be problematic? Any workarounds?

@Jarritus

Did you ever get 3D assets working?

@ArtFreex I am having issues with brushes as well. Did you find a solution?

I have some ability to modify the code so if you can point me in the right direction I can probably take care of it.

@ProtoNoob sadly no... I kinda gave up on getting those to work. And I also can't point you in the right direction as I can't seem to make sense of that error..