diff --git a/cc_pseudo_crawl/language_annotation/python_scripts/detect_html_lang_attrib.py b/cc_pseudo_crawl/language_annotation/python_scripts/detect_html_lang_attrib.py
index e69de29b..8d3cf9ac 100644
--- a/cc_pseudo_crawl/language_annotation/python_scripts/detect_html_lang_attrib.py
+++ b/cc_pseudo_crawl/language_annotation/python_scripts/detect_html_lang_attrib.py
@@ -0,0 +1,105 @@
+import os
+import logging
+import subprocess
+import sys
+from argparse import ArgumentParser
+from pathlib import Path
+
+import datasets
+import lxml.html
+from datasets import config, load_from_disk
+from datasets.utils.logging import set_verbosity_info
+
+set_verbosity_info()
+logger = logging.getLogger(__name__)
+
+# For `soup.decode_content` that can hit the limit
+sys.setrecursionlimit(10000)
+
+
+def get_args():
+    parser = ArgumentParser()
+    parser.add_argument(
+        "--dataset-path",
+        type=str,
+        required=True,
+        help="path to the parquet dataset folder",
+    )
+    parser.add_argument("--save-path", type=str, help="Where to save the datasets.")
+    parser.add_argument("--use-datasets-caching", action="store_true")
+    parser.add_argument(
+        "--num-proc", type=int, default=1, help="Number of procs use for preprocessing."
+    )
+    parser.add_argument(
+        "--num-examples",
+        type=int,
+        default=None,
+        help="Optional argument to select a subset (used for debugging purposes). Example `10`.",
+    )
+    args = parser.parse_args()
+
+    return args
+
+
+def main():
+    # Setup logging
+    logging.basicConfig(
+        format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
+        datefmt="%m/%d/%Y %H:%M:%S",
+        level=logging.INFO,
+    )
+    args = get_args()
+    logger.info(
+        f"** The job is runned with the following arguments: **\n{args}\n **** "
+    )
+
+    if os.path.isdir(args.save_path):
+        logger.info(f"Seed id {args.save_path.split('/')[-1]} already processed")
+        return
+
+    if not args.use_datasets_caching:
+        datasets.set_caching_enabled(False)
+    else:
+        logger.info(
+            f"the datasets results will be cached at {config.HF_DATASETS_CACHE}."
+        )
+
+    ds = load_from_disk(args.dataset_path)
+    logger.info(f"the dataset is {ds}")
+
+    if args.num_examples:
+        ds = ds.select([i for i in range(args.num_examples)])
+
+    def detect_lang(example):
+        if example["text"] is None or len(example["text"]) == 0:
+            example["html_lang_attr"] = None
+        else:
+            root = lxml.html.fromstring(example["html_str"])
+            root_lang = root.attrib.get("lang")
+            example["html_lang_attr"] = root_lang
+        return example
+
+    ds = ds.map(
+        detect_lang,
+        batched=False,
+        num_proc=args.num_proc,
+    )
+
+    if args.save_path:
+        save_path = Path(args.save_path)
+    else:
+        save_path = Path(args.dataset_path)
+
+    logger.info(
+        f"Lang attribute detected for {len([e for e in ds['train']['html_lang_attr'] if e is not None])} rows out of {len(ds['train'])} rows."
+    )
+
+    save_path_tmp = f"{str(save_path.absolute())}.tmp"
+    logger.info(f"Saving the dataset at {save_path_tmp}")
+    ds.save_to_disk(save_path_tmp)
+    logger.info(f"Moving the saved dataset to {str(save_path.absolute())}")
+    subprocess.run(["mv", save_path_tmp, str(save_path.absolute())])
+
+
+if __name__ == "__main__":
+    main()
diff --git a/cc_pseudo_crawl/language_annotation/slurm_scripts/02_detect_html_lang_attrib.slurm b/cc_pseudo_crawl/language_annotation/slurm_scripts/02_detect_html_lang_attrib.slurm
index e69de29b..02f250da 100644
--- a/cc_pseudo_crawl/language_annotation/slurm_scripts/02_detect_html_lang_attrib.slurm
+++ b/cc_pseudo_crawl/language_annotation/slurm_scripts/02_detect_html_lang_attrib.slurm
@@ -0,0 +1,35 @@
+#!/bin/bash
+#SBATCH --job-name=pseudo_crawl_extract_lang_tag
+#SBATCH --nodes=1
+#SBATCH --ntasks-per-node=1          # crucial - only 1 task per dist per node!
+#SBATCH --cpus-per-task=40           # number of cores per tasks
+#SBATCH --hint=nomultithread         # we get physical cores not logical
+#SBATCH --partition=cpu_p1
+#SBATCH --time 20:00:00              # maximum execution time (HH:MM:SS)
+#SBATCH --output=/gpfsscratch/rech/six/commun/pseudo_crawl/seeds_batch_2/logs/extract_lang_tag_V5/%x-%j.out           # output file name #TODO change path if necessary
+#SBATCH --array=35,341,297
+#SBATCH --account=six@cpu
+
+set -x -e
+
+source $six_ALL_CCFRWORK/start-prod
+conda activate thomas_data_tooling
+
+DATA_TOOLING_REPO=$WORK/repos/sync_data_tooling/data_tooling
+
+DATASET_PATH=/gpfsscratch/rech/six/urd43gx/crawl/shards/shard_"$SLURM_ARRAY_TASK_ID"
+SAVE_DATASET_DIR=$six_ALL_CCFRSCRATCH/pseudo_crawl/fasttext_annotation/seeds_batch_1/datasets-lang-annotation/bigscience-catalogue-data
+SAVE_DATASET_PATH=$SAVE_DATASET_DIR/seed_id="$SLURM_ARRAY_TASK_ID"
+echo $DATASET_PATH
+pushd $DATA_TOOLING_REPO
+
+mkdir -p $SAVE_DATASET_DIR
+
+export HF_DATASETS_OFFLINE=1
+export HF_DATASETS_CACHE=$SCRATCH/to_delete
+
+python -m cc_pseudo_crawl.language_annotation.python_scripts.detect_html_lang_attrib \
+    --dataset-path $DATASET_PATH \
+    --num-proc 40 \
+    --save-path $SAVE_DATASET_PATH \
+    --use-datasets-caching