diff --git a/backend/services/model_health_service.py b/backend/services/model_health_service.py
index 35fff2a23..5d472799d 100644
--- a/backend/services/model_health_service.py
+++ b/backend/services/model_health_service.py
@@ -38,8 +38,17 @@ def _normalize_embedding_url(base_url: str) -> str:
 def _infer_model_factory(model_type: str, base_url: str, current_factory: Optional[str] = None) -> Optional[str]:
     """Infer model_factory from base_url if not already set or is generic.
 
-    Uses the shared W11 host map so embedding and LLM/VLM inference do not drift.
+    For embedding/multi_embedding, uses legacy logic (only dashscope) to avoid
+    changing existing behavior. For other types (VLM), uses extended inference
+    so tokenpony URLs can be recognized for catalog healthcheck.
     """
+    # Embedding types: keep legacy behavior (only dashscope)
+    if model_type in EMBEDDING_TYPES:
+        if "dashscope" in base_url.lower():
+            return DASHSCOPE_MODEL_FACTORY
+        return current_factory
+
+    # Non-embedding types (VLM, etc): use extended inference
     try:
         from services.model_capacity_suggestion_service import pick_provider_from_base_url
 
diff --git a/scripts/generate_backfill_sql.py b/deploy/sql/migrations/generate_backfill_sql.py
similarity index 97%
rename from scripts/generate_backfill_sql.py
rename to deploy/sql/migrations/generate_backfill_sql.py
index 586c27e3d..239a5c93a 100644
--- a/scripts/generate_backfill_sql.py
+++ b/deploy/sql/migrations/generate_backfill_sql.py
@@ -2,7 +2,7 @@
 """Generate idempotent backfill SQL from capability_profiles.CATALOG.
 
 Usage:
-    python scripts/generate_backfill_sql.py > docker/sql/v2.2.x_MMDD_backfill_from_catalog.sql
+    python deploy/sql/migrations/generate_backfill_sql.py > deploy/sql/migrations/v2.2.x_MMDD_backfill_from_catalog.sql
 
 Run whenever capability_profiles.py changes, then commit the generated SQL.
 """
@@ -12,7 +12,7 @@
 from datetime import date
 from collections import namedtuple
 
-_project_root = os.path.join(os.path.dirname(__file__), "..")
+_project_root = os.path.join(os.path.dirname(__file__), "..", "..", "..")
 sys.path.insert(0, os.path.join(_project_root, "backend"))
 
 # Stub SDK types to avoid pulling in the full nexent SDK dependency chain
@@ -80,7 +80,7 @@ def main() -> None:
     today = date.today().strftime("%Y-%m-%d")
     lines: list[str] = []
 
-    lines.append(f"-- Generated by scripts/generate_backfill_sql.py on {today}")
+    lines.append(f"-- Generated by deploy/sql/migrations/generate_backfill_sql.py on {today}")
     lines.append(f"-- Catalog revision: {CATALOG_REVISION}")
     lines.append(f"-- Catalog entries: {len(CATALOG)}")
     lines.append("--")
diff --git a/frontend/app/[locale]/models/components/model/ModelAddDialog.tsx b/frontend/app/[locale]/models/components/model/ModelAddDialog.tsx
index 094ed4391..1dc6baafd 100644
--- a/frontend/app/[locale]/models/components/model/ModelAddDialog.tsx
+++ b/frontend/app/[locale]/models/components/model/ModelAddDialog.tsx
@@ -1171,7 +1171,6 @@ export const ModelAddDialog = ({
           apiKey: form.apiKey.trim() === "" ? "sk-no-api-key" : form.apiKey,
           maxTokens: maxTokensValue,
           displayName: form.displayName || form.name,
-          modelFactory: form.provider,
           ...(supportsCapacityFields ? buildCapacityPayload(form) : {}),
           ...acceptSignalKwargs,
         };
@@ -1214,7 +1213,6 @@ export const ModelAddDialog = ({
           apiKey: form.apiKey.trim() === "" ? "sk-no-api-key" : form.apiKey,
           maxTokens: maxTokensValue,
           displayName: form.displayName || form.name,
-          modelFactory: form.provider,
           ...(supportsCapacityFields ? buildCapacityPayload(form) : {}),
           ...acceptSignalKwargs,
         };