diff --git a/tests/e2e/test_legacy_ocr_backfill_smoke.py b/tests/e2e/test_legacy_ocr_backfill_smoke.py
new file mode 100644
index 00000000..b14dff87
--- /dev/null
+++ b/tests/e2e/test_legacy_ocr_backfill_smoke.py
@@ -0,0 +1,115 @@
+"""E2E smoke tests for legacy OCR backfill against real vault data.
+
+These tests require a real PaperForge vault with legacy OCR directories
+(pre-structured-pipeline, no version state in meta.json).
+
+Run:
+    pytest tests/e2e/test_legacy_ocr_backfill_smoke.py -v --tb=long
+
+Set PAPERFORGE_VAULT env var to override the default vault path.
+"""
+
+from __future__ import annotations
+
+import json
+import os
+from pathlib import Path
+
+import pytest
+
+_SMOKE_VAULT = os.environ.get("PAPERFORGE_VAULT", "")
+_VAULT_AVAILABLE = bool(_SMOKE_VAULT) and Path(_SMOKE_VAULT).exists()
+
+
+def _find_legacy_papers(vault: Path) -> list[dict]:
+    """Scan vault OCR dir for legacy papers (no version state, ocr_status=done)."""
+    from paperforge.worker.ocr_versions import classify_legacy_ocr_state
+    from paperforge.core.io import read_json
+
+    cfg_path = vault / "paperforge.json"
+    if not cfg_path.exists():
+        return []
+    cfg = json.loads(cfg_path.read_text(encoding="utf-8"))
+    system_dir = cfg.get("vault_config", {}).get("system_dir", "99_System")
+    ocr_root = vault / system_dir / "PaperForge" / "ocr"
+    if not ocr_root.exists():
+        return []
+
+    legacy = []
+    for paper_dir in ocr_root.iterdir():
+        if not paper_dir.is_dir():
+            continue
+        meta_path = paper_dir / "meta.json"
+        if not meta_path.exists():
+            continue
+        meta = read_json(meta_path)
+        state = classify_legacy_ocr_state(meta, ocr_dir=paper_dir)
+        if state["is_legacy"]:
+            legacy.append({
+                "key": paper_dir.name,
+                "can_backfill": state["can_backfill"],
+                "has_result_json": state["has_result_json"],
+            })
+    return legacy
+
+
+@pytest.mark.skipif(not _VAULT_AVAILABLE, reason="PAPERFORGE_VAULT not set or not found")
+def test_smoke_legacy_backfill_identifies_candidates():
+    """Verify that the vault has legacy OCR papers and they are correctly classified."""
+    vault = Path(_SMOKE_VAULT)
+    legacy = _find_legacy_papers(vault)
+
+    assert len(legacy) > 0, (
+        f"No legacy OCR papers found in {vault}. "
+        "This is expected if all papers have already been backfilled or if "
+        "the vault uses the structured pipeline exclusively."
+    )
+
+    # At least some should be backfillable (have result.json)
+    backfillable = [p for p in legacy if p["can_backfill"]]
+    if not backfillable:
+        pytest.skip("No backfillable legacy papers found (no result.json available)")
+
+
+@pytest.mark.skipif(not _VAULT_AVAILABLE, reason="PAPERFORGE_VAULT not set or not found")
+def test_smoke_backfill_single_legacy_paper(tmp_path):
+    """Backfill a single legacy paper into a temp vault clone and verify artifacts."""
+    import shutil
+
+    from paperforge.worker.ocr_rebuild import backfill_from_result, select_legacy_papers_for_backfill
+
+    vault = Path(_SMOKE_VAULT)
+    legacy = [p for p in _find_legacy_papers(vault) if p["can_backfill"]]
+    if not legacy:
+        pytest.skip("No backfillable legacy papers for smoke test")
+
+    target = legacy[0]
+    paper_dir = vault / _get_system_dir(vault) / "PaperForge" / "ocr" / target["key"]
+
+    # Copy this paper's OCR dir to temp vault
+    temp_vault = tmp_path / "vault"
+    shutil.copytree(vault / "paperforge.json", temp_vault / "paperforge.json")
+    temp_ocr_dest = temp_vault / _get_system_dir(vault) / "PaperForge" / "ocr" / target["key"]
+    shutil.copytree(paper_dir, temp_ocr_dest)
+
+    # Run backfill
+    result = backfill_from_result(temp_vault, target["key"])
+
+    assert result["backfill_status"] == "done", (
+        f"Backfill failed: {result.get('error', 'unknown error')}"
+    )
+
+    # Verify derived artifacts
+    assert (temp_ocr_dest / "index" / "role-index.json").exists()
+    assert (temp_ocr_dest / "health" / "ocr_health.json").exists()
+    assert (temp_ocr_dest / "render" / "fulltext.md").exists()
+
+    # Verify backfill metadata
+    meta = json.loads((temp_ocr_dest / "meta.json").read_text(encoding="utf-8"))
+    assert meta.get("is_backfilled") is True
+    assert "backfilled_at" in meta
+
+
+def _get_system_dir(vault: Path) -> str:
+    cfg = json.loads((vault / "paperforge.json").read_text(encoding="utf-8"))
+    return cfg.get("vault_config", {}).get("system_dir", "99_System")