diff --git a/src/grok_search/server.py b/src/grok_search/server.py
index 7754216..d46a5ae 100644
--- a/src/grok_search/server.py
+++ b/src/grok_search/server.py
@@ -15,13 +15,13 @@
     from grok_search.providers.grok import GrokSearchProvider
     from grok_search.logger import log_info
     from grok_search.config import config
-    from grok_search.sources import SourcesCache, merge_sources, new_session_id, split_answer_and_sources
+    from grok_search.sources import SourcesCache, merge_sources, new_session_id, sanitize_answer_text, split_answer_and_sources
     from grok_search.planning import engine as planning_engine, _split_csv
 except ImportError:
     from .providers.grok import GrokSearchProvider
     from .logger import log_info
     from .config import config
-    from .sources import SourcesCache, merge_sources, new_session_id, split_answer_and_sources
+    from .sources import SourcesCache, merge_sources, new_session_id, sanitize_answer_text, split_answer_and_sources
     from .planning import engine as planning_engine, _split_csv
 
 import asyncio
@@ -203,6 +203,7 @@ async def _safe_firecrawl() -> list[dict] | None:
         firecrawl_results = gathered[idx]
 
     answer, grok_sources = split_answer_and_sources(grok_result)
+    answer = sanitize_answer_text(answer)
     extra = _extra_results_to_sources(tavily_results, firecrawl_results)
     all_sources = merge_sources(grok_sources, extra)
 
diff --git a/src/grok_search/sources.py b/src/grok_search/sources.py
index 63386e2..6df7ad1 100644
--- a/src/grok_search/sources.py
+++ b/src/grok_search/sources.py
@@ -23,6 +23,10 @@
 _SOURCES_FUNCTION_PATTERN = re.compile(
     r"(?im)(^|\n)\s*(sources|source|citations|citation|references|reference|citation_card|source_cards|source_card)\s*\("
 )
+_THINK_TAG_PATTERN = re.compile(
+    r"<think\b[^>]*>.*?</think>",
+    re.IGNORECASE | re.DOTALL,
+)
 
 
 def new_session_id() -> str:
@@ -67,6 +71,16 @@ def merge_sources(*source_lists: list[dict]) -> list[dict]:
     return merged
 
 
+def sanitize_answer_text(text: str) -> str:
+    """Remove model reasoning tags from answer text while preserving content."""
+    raw = (text or "").strip()
+    if not raw:
+        return ""
+    cleaned = _THINK_TAG_PATTERN.sub("", raw)
+    cleaned = re.sub(r"\n{3,}", "\n\n", cleaned).strip()
+    return cleaned
+
+
 def split_answer_and_sources(text: str) -> tuple[str, list[dict]]:
     raw = (text or "").strip()
     if not raw: