diff --git a/backend/main.py b/backend/main.py
index adba9583b8..910a059d8a 100644
--- a/backend/main.py
+++ b/backend/main.py
@@ -3,6 +3,7 @@
 
 import firebase_admin
 from fastapi import FastAPI
+import utils.retrieval.graph as graph
 
 from modal import Image, App, asgi_app, Secret, Cron
 from routers import workflow, chat, firmware, plugins, memories, transcribe_v2, notifications, \
diff --git a/backend/requirements.txt b/backend/requirements.txt
index dcf66e04c6..54f1ef2ab6 100644
--- a/backend/requirements.txt
+++ b/backend/requirements.txt
@@ -1,7 +1,7 @@
 aenum==3.1.15
 aiofiles==24.1.0
 aiohappyeyeballs==2.3.4
-aiohttp==3.10.1
+aiohttp==3.9.5
 aiosignal==1.3.1
 aiostream==0.5.2
 alembic==1.13.2
@@ -61,6 +61,7 @@ google-cloud-storage==2.18.0
 google-crc32c==1.5.0
 google-resumable-media==2.7.1
 googleapis-common-protos==1.63.2
+graphviz==0.20.3
 groq==0.9.0
 grpcio==1.65.4
 grpcio-status==1.62.3
@@ -82,6 +83,7 @@ idna==3.7
 ipython==8.26.0
 jedi==0.19.1
 Jinja2==3.1.4
+jiter==0.6.1
 jiwer==3.0.4
 joblib==1.4.2
 jsonpatch==1.33
@@ -90,14 +92,17 @@ jsonschema==4.23.0
 jsonschema-specifications==2023.12.1
 julius==0.2.7
 kiwisolver==1.4.5
-langchain==0.2.12
-langchain-community==0.2.11
-langchain-core==0.2.28
+langchain==0.3.4
+langchain-community==0.3.3
+langchain-core==0.3.12
 langchain-groq==0.1.9
-langchain-openai==0.1.20
-langchain-pinecone==0.1.3
-langchain-text-splitters==0.2.2
-langsmith==0.1.96
+langchain-openai==0.2.3
+langchain-pinecone==0.2.0
+langchain-text-splitters==0.3.0
+langgraph==0.2.39
+langgraph-checkpoint==2.0.1
+langgraph-sdk==0.1.33
+langsmith==0.1.137
 lazy_loader==0.4
 librosa==0.10.2.post1
 lightning==2.4.0
@@ -116,7 +121,7 @@ more-itertools==10.5.0
 mplcursors==0.5.3
 mpld3==0.5.10
 mpmath==1.3.0
-msgpack==1.0.8
+msgpack==1.1.0
 multidict==6.0.5
 mypy-extensions==1.0.0
 narwhals==1.5.2
@@ -127,7 +132,7 @@ numba==0.60.0
 numpy==1.26.4
 omegaconf==2.3.0
 onnxruntime==1.19.0
-openai==1.39.0
+openai==1.52.2
 optuna==3.6.1
 opuslib==3.0.1
 orjson==3.10.6
@@ -161,13 +166,15 @@ pyasn1==0.6.0
 pyasn1_modules==0.4.0
 pycparser==2.22
 pydantic==2.8.2
+pydantic-settings==2.6.0
 pydantic_core==2.20.1
 pydeck==0.9.1
 pydub==0.25.1
 Pygments==2.18.0
+pygraphviz==1.14
 PyJWT==2.9.0
 pynndescent==0.5.13
-PyOgg@ git+https://github.com/TeamPyOgg/PyOgg@6871a4f234e8a3a346c4874a12509bfa02c4c63a
+PyOgg @ git+https://github.com/TeamPyOgg/PyOgg@6871a4f234e8a3a346c4874a12509bfa02c4c63a
 pyparsing==3.1.2
 python-dateutil==2.9.0.post0
 python-dotenv==1.0.1
@@ -182,6 +189,7 @@ redis==5.0.8
 referencing==0.35.1
 regex==2024.7.24
 requests==2.32.3
+requests-toolbelt==1.0.0
 rich==13.7.1
 rpds-py==0.20.0
 rsa==4.9
@@ -241,4 +249,4 @@ watchfiles==0.22.0
 wcwidth==0.2.13
 webrtcvad==2.0.10
 websockets==12.0
-yarl==1.9.4
\ No newline at end of file
+yarl==1.9.4
diff --git a/backend/scripts/users/retrieval.py b/backend/scripts/users/retrieval.py
index 5dd895ad68..b527b455df 100644
--- a/backend/scripts/users/retrieval.py
+++ b/backend/scripts/users/retrieval.py
@@ -44,7 +44,8 @@ class ExtractedInformation(BaseModel):
     )
 
 
-def migrate_memory_vector_metadata(memory_id: str, created_at: datetime, transcript_segment: List[dict]) -> ExtractedInformation:
+def migrate_memory_vector_metadata(memory_id: str, created_at: datetime,
+                                   transcript_segment: List[dict]) -> ExtractedInformation:
     transcript = ''
     for segment in transcript_segment:
         transcript += f'{segment["text"].strip()}\n\n'
@@ -82,6 +83,8 @@ def migrate_memory_vector_metadata(memory_id: str, created_at: datetime, transcr
 
 
 if __name__ == '__main__':
+    # TODO: finish migration script
+    # TODO: inlude process_memory to process this too
     uids = get_users_uid()
     for uid in ['TtCJi59JTVXHmyUC6vUQ1d9U6cK2']:
         memories = memories_db.get_memories(uid, limit=1000)
diff --git a/backend/utils/retrieval/graph.py b/backend/utils/retrieval/graph.py
new file mode 100644
index 0000000000..d476bdc1c1
--- /dev/null
+++ b/backend/utils/retrieval/graph.py
@@ -0,0 +1,87 @@
+from typing import Literal
+
+from langchain_openai import ChatOpenAI
+from langgraph.checkpoint.memory import MemorySaver
+from langgraph.constants import END
+from langgraph.graph import START, StateGraph, MessagesState
+
+model = ChatOpenAI(model='gpt-4o-mini')
+
+
+def determine_conversation_type(s: MessagesState) -> Literal["no_context_conversation", "context_dependent_conversation"]:
+    # call requires context
+    # if requires context, spawn 2 parallel graphs edges?
+    return 'no_context_conversation'
+
+
+def no_context_conversation(state: MessagesState):
+    # continue the conversation
+    return END
+
+
+def context_dependent_conversation(state: MessagesState):
+    pass
+
+
+# TODO: include a question extractor? node?
+
+
+def retrieve_topics_filters(state: MessagesState):
+    # retrieve all available entities, names, topics, etc, and ask it to filter based on the question.
+    return 'query_vectors'
+
+
+def retrieve_date_filters(state: MessagesState):
+    # extract dates filters, and send them to qa_handler node
+    return 'query_vectors'
+
+
+def query_vectors(state: MessagesState):
+    # receives both filters, and finds vectors + rerank them
+    # TODO: maybe didnt find anything, tries RAG, or goes to simple conversation?
+    pass
+
+
+def qa_handler(state: MessagesState):
+    # takes vectors found, retrieves memories, and does QA on them
+    return END
+
+
+workflow = StateGraph(MessagesState)  # custom state?
+
+
+
+
+workflow.add_edge(START, "determine_conversation_type")
+workflow.add_node('determine_conversation_type', determine_conversation_type)
+
+workflow.add_conditional_edges(
+    "determine_conversation_type",
+    determine_conversation_type,
+)
+
+workflow.add_node("no_context_conversation", no_context_conversation)
+workflow.add_node("context_dependent_conversation", context_dependent_conversation)
+
+workflow.add_edge("context_dependent_conversation", "retrieve_topics_filters")
+workflow.add_edge("context_dependent_conversation", "retrieve_date_filters")
+
+workflow.add_node("retrieve_topics_filters", retrieve_topics_filters)
+workflow.add_node("retrieve_date_filters", retrieve_date_filters)
+
+workflow.add_edge('retrieve_topics_filters', 'query_vectors')
+workflow.add_edge('retrieve_date_filters', 'query_vectors')
+
+workflow.add_node('query_vectors', query_vectors)
+
+workflow.add_edge('query_vectors', 'qa_handler')
+
+workflow.add_node('qa_handler', qa_handler)
+
+workflow.add_edge('qa_handler', END)
+
+checkpointer = MemorySaver()
+graph = workflow.compile(checkpointer=checkpointer)
+
+if __name__ == '__main__':
+    graph.get_graph().draw_png('workflow.png')