Allow created_at override on /memories/raw

API.md

@@ -46,7 +46,7 @@ Returns the mem0 add response with the created memory payload.
 ---
 
 ### `POST /memories/raw`
-Store a pre-processed or pre-summarised memory directly in the conversational collection without running the mem0 extraction LLM. Works just like `/knowledge` in that `infer` is forced to `false` and metadata is stored verbatim, making it a fast way to inject cleaned data. You can also set `metadata.created_at` (ISO 8601 string) to control the stored timestamp; otherwise the server uses the current time.
+Store a pre-processed or pre-summarised memory directly in the conversational collection without running the mem0 extraction LLM. Works just like `/knowledge` in that `infer` is forced to `false` and metadata is stored verbatim, making it a fast way to inject cleaned data. If you set `metadata.created_at` (ISO 8601 string), the server overrides the mem0 timestamp for this route; otherwise it uses the current time.
 
 **Request**
 ```json

mem0core/handlers.py

@@ -19,17 +19,56 @@ def sanitize_metadata(meta: dict) -> dict:
     return clean
 
 
+def extract_ids(result: dict | list | None) -> list[str]:
+    """Best-effort extraction of memory IDs from a mem0 add() response."""
+    if not result:
+        return []
+    if isinstance(result, dict):
+        if isinstance(result.get("id"), str):
+            return [result["id"]]
+        if isinstance(result.get("memory_id"), str):
+            return [result["memory_id"]]
+        results = result.get("results") or result.get("memories") or []
+        if isinstance(results, list):
+            return [r.get("id") for r in results if isinstance(r, dict) and r.get("id")]
+    if isinstance(result, list):
+        return [r.get("id") for r in result if isinstance(r, dict) and r.get("id")]
+    return []
+
+
+def override_created_at(mem: Memory, memory_ids: list[str], created_at: str) -> None:
+    """Force a created_at override in Chroma metadata for the given memories."""
+    if not memory_ids or not created_at:
+        return
+    collection = mem.vector_store.collection
+    existing = collection.get(ids=memory_ids, include=["metadatas"])
+    metadatas = existing.get("metadatas", [])
+    updated = []
+    for meta in metadatas:
+        meta = meta or {}
+        meta["created_at"] = created_at
+        updated.append(meta)
+    if updated:
+        collection.update(ids=memory_ids, metadatas=updated)
+
+
 def extract_user_id(data: dict) -> str:
     """Read user ID from either snake_case or camelCase payload fields."""
     return data.get("userId") or data.get("user_id") or DEFAULT_USER_ID
 
 
-async def handle_add(req: Request, mem: Memory, verbatim_allowed: bool = False):
+async def handle_add(
+    req: Request,
+    mem: Memory,
+    verbatim_allowed: bool = False,
+    allow_created_at_override: bool = False,
+):
     """Handle add for /memories and /knowledge with optional verbatim mode."""
     data = await req.json()
     user_id = extract_user_id(data)
     raw_meta = data.get("metadata")
     metadata = sanitize_metadata(raw_meta) if raw_meta else None
+    desired_created_at = metadata.get("created_at") if metadata else None
     messages = data.get("messages")
     text = data.get("text")
 
@@ -39,6 +78,11 @@ async def handle_add(req: Request, mem: Memory, verbatim_allowed: bool = False):
     if verbatim_allowed:
         content = text or " ".join(m["content"] for m in messages if m.get("role") == "user")
         result = mem.add(content, user_id=user_id, metadata=metadata, infer=False)
+        if allow_created_at_override and desired_created_at:
+            memory_ids = extract_ids(result)
+            override_created_at(mem, memory_ids, desired_created_at)
+            if isinstance(result, dict) and metadata:
+                result["metadata"] = {**(result.get("metadata") or {}), "created_at": desired_created_at}
         print(f"[add verbatim] user={user_id} chars={len(content)} meta={metadata}")
         return SafeJSONResponse(content=result)
 

mem0core/routes.py

@@ -48,7 +48,12 @@ def build_router(memory_conv: Memory, memory_know: Memory) -> APIRouter:
     @router.post("/memories/raw", summary="Store raw conversational memory", tags=["memories"])
     async def add_raw_memory(req: Request):
         """Store processed or pre-summarized conversational memory without rerunning the mem0 LLM."""
-        return await handle_add(req, memory_conv, verbatim_allowed=True)
+        return await handle_add(
+            req,
+            memory_conv,
+            verbatim_allowed=True,
+            allow_created_at_override=True,
+        )
 
     @router.post("/memories/search", summary="Search conversational memory", tags=["memories"])
     async def search_memories(req: Request):