diff --git a/zetta_utils/cli/main.py b/zetta_utils/cli/main.py
index 04cfc41fb..9ccd01c7b 100644
--- a/zetta_utils/cli/main.py
+++ b/zetta_utils/cli/main.py
@@ -7,7 +7,7 @@
 import zetta_utils
 from zetta_utils import log
 from zetta_utils.parsing import json
-from zetta_utils.run import run_ctx_manager
+from zetta_utils.run import record_run, run_ctx_manager
 
 logger = log.get_logger("zetta_utils")
 
@@ -125,6 +125,7 @@ def run(
     if heartbeat is False:
         run_ctx = run_ctx_manager(run_id=run_id, heartbeat_interval=-1)
     with run_ctx:
+        record_run(path)
         result = zetta_utils.builder.build(spec, parallel=parallel_builder)
         logger.debug(f"Outcome: {pprint.pformat(result, indent=4)}")
         if pdb:
diff --git a/zetta_utils/run/__init__.py b/zetta_utils/run/__init__.py
index 4e024c71b..9c580b536 100644
--- a/zetta_utils/run/__init__.py
+++ b/zetta_utils/run/__init__.py
@@ -6,6 +6,8 @@
 from typing import Optional
 
 import attrs
+import fsspec
+from cloudfiles import paths
 
 from zetta_utils import log
 from zetta_utils.common import RepeatTimer
@@ -54,6 +56,28 @@ def register_clusters(clusters: list) -> None:  # pragma: no cover
     _update_run_info(info)
 
 
+def record_run(spec_path: str | None = None) -> None:  # pragma: no cover
+    """
+    Records run info in a bucket for archiving.
+    """
+    zetta_user = os.environ["ZETTA_USER"]
+    info_path = os.environ.get("RUN_INFO_BUCKET", RUN_INFO_BUCKET)
+    info_path_user = os.path.join(info_path, zetta_user)
+    run_info = {
+        "zetta_user": zetta_user,
+        "zetta_project": os.environ["ZETTA_PROJECT"],
+        "json_spec": json.loads(os.environ["ZETTA_RUN_SPEC"]),
+    }
+    with fsspec.open(os.path.join(info_path_user, f"{RUN_ID}.json"), "w") as f:
+        json.dump(run_info, f, indent=2)
+
+    if spec_path is not None:
+        with open(spec_path, "r", encoding="utf-8") as src:
+            content = src.read()
+            with fsspec.open(os.path.join(info_path_user, f"{RUN_ID}.cue"), "w") as dst:
+                dst.write(content)
+
+
 def _update_run_info(info: DBRowDataT) -> None:  # pragma: no cover
     row_key = f"run-{RUN_ID}"
     col_keys = tuple(info.keys())