diff --git a/backend/src/plugins/translate/_llm_http.py b/backend/src/plugins/translate/_llm_http.py
index 869215da..d3127edd 100644
--- a/backend/src/plugins/translate/_llm_http.py
+++ b/backend/src/plugins/translate/_llm_http.py
@@ -11,12 +11,19 @@
 # @RATIONALE Extracted from LLMTranslationService (793 lines) to keep module under INV_7 limit.
 # @REJECTED Single HTTP client class — kept as module-level functions for stateless reusability.
 
+import os
 import time
 from typing import Any
 
 from ...core.logger import logger
 
 
+def _get_verify() -> bool:
+    """Resolve SSL verify flag from LLM_SSL_VERIFY env var."""
+    raw = os.getenv("LLM_SSL_VERIFY", "true").strip().lower()
+    return raw not in ("false", "0", "no", "off")
+
+
 # #region call_openai_compatible [C:3] [TYPE Function] [SEMANTICS translate, llm, http, openai]
 # @BRIEF Call OpenAI-compatible API with rate-limit handling and structured output fallback.
 # @PRE Valid API endpoint, key, model, and prompt.
@@ -141,7 +148,7 @@ def _do_http_request(url: str, headers: dict, payload: dict) -> tuple[Any, str]:
     _max_retry_429 = 3
     _retry_count_429 = 0
     while _retry_count_429 < _max_retry_429:
-        response = http_requests.post(url, headers=headers, json=payload, timeout=180)
+        response = http_requests.post(url, headers=headers, json=payload, timeout=180, verify=_get_verify())
         response_text = response.text
         if response.status_code == 429:
             _retry_count_429 += 1
@@ -179,7 +186,7 @@ def _handle_response_format_fallback(
         logger.explore("Structured outputs not supported, retrying without response_format",
                        extra={"src": "executor"})
         payload.pop("response_format", None)
-        new_response = http_requests.post(url, headers=headers, json=payload, timeout=180)
+        new_response = http_requests.post(url, headers=headers, json=payload, timeout=180, verify=_get_verify())
         response.status_code = new_response.status_code
         response._content = new_response.content
         response.encoding = new_response.encoding
diff --git a/backend/src/plugins/translate/preview_llm_client.py b/backend/src/plugins/translate/preview_llm_client.py
index cad17984..e3fd58a8 100644
--- a/backend/src/plugins/translate/preview_llm_client.py
+++ b/backend/src/plugins/translate/preview_llm_client.py
@@ -3,12 +3,19 @@
 # @SIDE_EFFECT Makes HTTP POST calls to external LLM API.
 # @RELATION DEPENDS_ON -> [EXT:requests]
 
+import os
 import time as _time
 from typing import Any
 
 from ...core.logger import logger
 
 
+def _get_verify() -> bool:
+    """Resolve SSL verify flag from LLM_SSL_VERIFY env var."""
+    raw = os.getenv("LLM_SSL_VERIFY", "true").strip().lower()
+    return raw not in ("false", "0", "no", "off")
+
+
 class LLMClient:
     """Call OpenAI-compatible LLM APIs with retry and structured output handling."""
 
@@ -63,7 +70,7 @@ class LLMClient:
         _max_retry_429 = 3
         _retry_count_429 = 0
         while _retry_count_429 < _max_retry_429:
-            response = http_requests.post(url, headers=headers, json=payload, timeout=600)
+            response = http_requests.post(url, headers=headers, json=payload, timeout=600, verify=_get_verify())
             if response.status_code == 429:
                 _retry_count_429 += 1
                 retry_after = response.headers.get("Retry-After")
@@ -80,7 +87,7 @@ class LLMClient:
                 and any(p in (response.text or "").lower() for p in _response_format_error_patterns)):
             logger.explore("Structured outputs not supported, retrying without response_format")
             payload.pop("response_format", None)
-            response = http_requests.post(url, headers=headers, json=payload, timeout=600)
+            response = http_requests.post(url, headers=headers, json=payload, timeout=600, verify=_get_verify())
 
         if not response.ok:
             logger.explore(f"LLM API error status={response.status_code} model={payload.get('model')} body={response.text[:2000]}")