diff --git a/.github/workflows/e2e_tests.yml b/.github/workflows/e2e_tests.yml
index 72f15e0da..e47dee417 100644
--- a/.github/workflows/e2e_tests.yml
+++ b/.github/workflows/e2e_tests.yml
@@ -30,6 +30,10 @@ on:
         required: true
       GRAPHISTRY_PASSWORD:
         required: true
+      AWS_ACCESS_KEY_ID:
+        required: true
+      AWS_SECRET_ACCESS_KEY:
+        required: true
 
 jobs:
   run-server-start-test:
@@ -157,3 +161,30 @@ jobs:
           EMBEDDING_API_KEY: ${{ secrets.EMBEDDING_API_KEY }}
           EMBEDDING_API_VERSION: ${{ secrets.EMBEDDING_API_VERSION }}
         run: poetry run python ./cognee/tests/test_deduplication.py
+
+  run-s3-bucket-test:
+      name: S3 Bucket Test
+      runs-on: ubuntu-22.04
+      steps:
+        - name: Check out repository
+          uses: actions/checkout@v4
+
+        - name: Cognee Setup
+          uses: ./.github/actions/cognee_setup
+          with:
+            python-version: '3.11.x'
+
+        - name: Run S3 Bucket Test
+          env:
+            ENV: 'dev'
+            LLM_MODEL: ${{ secrets.LLM_MODEL }}
+            LLM_ENDPOINT: ${{ secrets.LLM_ENDPOINT }}
+            LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
+            LLM_API_VERSION: ${{ secrets.LLM_API_VERSION }}
+            EMBEDDING_MODEL: ${{ secrets.EMBEDDING_MODEL }}
+            EMBEDDING_ENDPOINT: ${{ secrets.EMBEDDING_ENDPOINT }}
+            EMBEDDING_API_KEY: ${{ secrets.EMBEDDING_API_KEY }}
+            EMBEDDING_API_VERSION: ${{ secrets.EMBEDDING_API_VERSION }}
+            AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
+            AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+          run: poetry run python ./cognee/tests/test_s3.py
diff --git a/cognee/api/v1/add/config.py b/cognee/api/v1/add/config.py
new file mode 100644
index 000000000..dc08e3333
--- /dev/null
+++ b/cognee/api/v1/add/config.py
@@ -0,0 +1,14 @@
+from typing import Optional
+from functools import lru_cache
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class S3Config(BaseSettings):
+    aws_access_key_id: Optional[str] = None
+    aws_secret_access_key: Optional[str] = None
+    model_config = SettingsConfigDict(env_file=".env", extra="allow")
+
+
+@lru_cache
+def get_s3_config():
+    return S3Config()
diff --git a/cognee/infrastructure/databases/vector/embeddings/OllamaEmbeddingEngine.py b/cognee/infrastructure/databases/vector/embeddings/OllamaEmbeddingEngine.py
index 072ef9ac4..26b98df08 100644
--- a/cognee/infrastructure/databases/vector/embeddings/OllamaEmbeddingEngine.py
+++ b/cognee/infrastructure/databases/vector/embeddings/OllamaEmbeddingEngine.py
@@ -9,7 +9,7 @@ import aiohttp.http_exceptions
 from cognee.infrastructure.databases.vector.embeddings.EmbeddingEngine import EmbeddingEngine
 from cognee.infrastructure.databases.exceptions.EmbeddingException import EmbeddingException
 from cognee.infrastructure.llm.tokenizer.HuggingFace import HuggingFaceTokenizer
-from cognee.infrastructure.llm.rate_limiter import (
+from cognee.infrastructure.llm.embedding_rate_limiter import (
     embedding_rate_limit_async,
     embedding_sleep_and_retry_async,
 )
diff --git a/cognee/infrastructure/files/utils/get_file_metadata.py b/cognee/infrastructure/files/utils/get_file_metadata.py
index e5d884c60..aea34c04c 100644
--- a/cognee/infrastructure/files/utils/get_file_metadata.py
+++ b/cognee/infrastructure/files/utils/get_file_metadata.py
@@ -20,7 +20,7 @@ def get_file_metadata(file: BinaryIO) -> FileMetadata:
 
     file_type = guess_file_type(file)
 
-    file_path = file.name
+    file_path = getattr(file, "name", None) or getattr(file, "full_name", None)
     file_name = str(file_path).split("/")[-1].split(".")[0] if file_path else None
 
     return FileMetadata(
diff --git a/cognee/infrastructure/llm/openai/adapter.py b/cognee/infrastructure/llm/openai/adapter.py
index 881e2db62..417af85df 100644
--- a/cognee/infrastructure/llm/openai/adapter.py
+++ b/cognee/infrastructure/llm/openai/adapter.py
@@ -6,6 +6,8 @@ from typing import Type
 import litellm
 import instructor
 from pydantic import BaseModel
+
+from cognee.modules.data.processing.document_types.open_data_file import open_data_file
 from cognee.shared.data_models import MonitoringTool
 from cognee.exceptions import InvalidValueError
 from cognee.infrastructure.llm.llm_interface import LLMInterface
@@ -114,26 +116,24 @@ class OpenAIAdapter(LLMInterface):
     def create_transcript(self, input):
         """Generate a audio transcript from a user query."""
 
-        if not os.path.isfile(input):
+        if not input.startswith("s3://") and not os.path.isfile(input):
             raise FileNotFoundError(f"The file {input} does not exist.")
 
-        # with open(input, 'rb') as audio_file:
-        #     audio_data = audio_file.read()
-
-        transcription = litellm.transcription(
-            model=self.transcription_model,
-            file=Path(input),
-            api_key=self.api_key,
-            api_base=self.endpoint,
-            api_version=self.api_version,
-            max_retries=self.MAX_RETRIES,
-        )
+        with open_data_file(input, mode="rb") as audio_file:
+            transcription = litellm.transcription(
+                model=self.transcription_model,
+                file=audio_file,
+                api_key=self.api_key,
+                api_base=self.endpoint,
+                api_version=self.api_version,
+                max_retries=self.MAX_RETRIES,
+            )
 
         return transcription
 
     @rate_limit_sync
     def transcribe_image(self, input) -> BaseModel:
-        with open(input, "rb") as image_file:
+        with open_data_file(input, mode="rb") as image_file:
             encoded_image = base64.b64encode(image_file.read()).decode("utf-8")
 
         return litellm.completion(
diff --git a/cognee/modules/data/processing/document_types/PdfDocument.py b/cognee/modules/data/processing/document_types/PdfDocument.py
index 12b919b4a..7ddbc0885 100644
--- a/cognee/modules/data/processing/document_types/PdfDocument.py
+++ b/cognee/modules/data/processing/document_types/PdfDocument.py
@@ -1,6 +1,6 @@
 from pypdf import PdfReader
 from cognee.modules.chunking.Chunker import Chunker
-
+from .open_data_file import open_data_file
 from .Document import Document
 
 
@@ -8,15 +8,14 @@ class PdfDocument(Document):
     type: str = "pdf"
 
     def read(self, chunker_cls: Chunker, max_chunk_size: int):
-        file = PdfReader(self.raw_data_location)
+        with open_data_file(self.raw_data_location, mode="rb") as stream:
+            file = PdfReader(stream)
 
-        def get_text():
-            for page in file.pages:
-                page_text = page.extract_text()
-                yield page_text
+            def get_text():
+                for page in file.pages:
+                    page_text = page.extract_text()
+                    yield page_text
 
-        chunker = chunker_cls(self, get_text=get_text, max_chunk_size=max_chunk_size)
+            chunker = chunker_cls(self, get_text=get_text, max_chunk_size=max_chunk_size)
 
-        yield from chunker.read()
-
-        file.stream.close()
+            yield from chunker.read()
diff --git a/cognee/modules/data/processing/document_types/TextDocument.py b/cognee/modules/data/processing/document_types/TextDocument.py
index 8a65ec05b..db6b6a842 100644
--- a/cognee/modules/data/processing/document_types/TextDocument.py
+++ b/cognee/modules/data/processing/document_types/TextDocument.py
@@ -1,5 +1,6 @@
 from .Document import Document
 from cognee.modules.chunking.Chunker import Chunker
+from .open_data_file import open_data_file
 
 
 class TextDocument(Document):
@@ -7,15 +8,12 @@ class TextDocument(Document):
 
     def read(self, chunker_cls: Chunker, max_chunk_size: int):
         def get_text():
-            with open(self.raw_data_location, mode="r", encoding="utf-8") as file:
+            with open_data_file(self.raw_data_location, mode="r", encoding="utf-8") as file:
                 while True:
                     text = file.read(1000000)
-
-                    if len(text.strip()) == 0:
+                    if not text.strip():
                         break
-
                     yield text
 
         chunker = chunker_cls(self, max_chunk_size=max_chunk_size, get_text=get_text)
-
         yield from chunker.read()
diff --git a/cognee/modules/data/processing/document_types/UnstructuredDocument.py b/cognee/modules/data/processing/document_types/UnstructuredDocument.py
index 731f558b4..b95b93fae 100644
--- a/cognee/modules/data/processing/document_types/UnstructuredDocument.py
+++ b/cognee/modules/data/processing/document_types/UnstructuredDocument.py
@@ -2,6 +2,7 @@ from io import StringIO
 
 from cognee.modules.chunking.Chunker import Chunker
 from cognee.modules.data.exceptions import UnstructuredLibraryImportError
+from cognee.modules.data.processing.document_types.open_data_file import open_data_file
 
 from .Document import Document
 
@@ -16,16 +17,19 @@ class UnstructuredDocument(Document):
             except ModuleNotFoundError:
                 raise UnstructuredLibraryImportError
 
-            elements = partition(self.raw_data_location, content_type=self.mime_type)
+            if self.raw_data_location.startswith("s3://"):
+                with open_data_file(self.raw_data_location, mode="rb") as f:
+                    elements = partition(file=f, content_type=self.mime_type)
+            else:
+                elements = partition(self.raw_data_location, content_type=self.mime_type)
+
             in_memory_file = StringIO("\n\n".join([str(el) for el in elements]))
             in_memory_file.seek(0)
 
             while True:
                 text = in_memory_file.read(1024)
-
-                if len(text.strip()) == 0:
+                if not text.strip():
                     break
-
                 yield text
 
         chunker = chunker_cls(self, get_text=get_text, max_chunk_size=max_chunk_size)
diff --git a/cognee/modules/data/processing/document_types/open_data_file.py b/cognee/modules/data/processing/document_types/open_data_file.py
new file mode 100644
index 000000000..207b67fdb
--- /dev/null
+++ b/cognee/modules/data/processing/document_types/open_data_file.py
@@ -0,0 +1,26 @@
+import s3fs
+from typing import IO, Optional
+from cognee.api.v1.add.config import get_s3_config
+
+
+def open_data_file(
+    file_path: str, mode: str = "rb", encoding: Optional[str] = None, **kwargs
+) -> IO:
+    if file_path.startswith("s3://"):
+        s3_config = get_s3_config()
+        if s3_config.aws_access_key_id is not None and s3_config.aws_secret_access_key is not None:
+            fs = s3fs.S3FileSystem(
+                key=s3_config.aws_access_key_id, secret=s3_config.aws_secret_access_key, anon=False
+            )
+        else:
+            raise ValueError("S3 credentials are not set in the configuration.")
+
+        if "b" in mode:
+            f = fs.open(file_path, mode=mode, **kwargs)
+            if not hasattr(f, "name") or not f.name:
+                f.name = file_path.split("/")[-1]
+            return f
+        else:
+            return fs.open(file_path, mode=mode, encoding=encoding, **kwargs)
+    else:
+        return open(file_path, mode=mode, encoding=encoding, **kwargs)
diff --git a/cognee/modules/ingestion/classify.py b/cognee/modules/ingestion/classify.py
index dd52df86a..5f3bbbd38 100644
--- a/cognee/modules/ingestion/classify.py
+++ b/cognee/modules/ingestion/classify.py
@@ -1,18 +1,22 @@
 from io import BufferedReader
-from typing import Union, BinaryIO
-from .data_types import TextData, BinaryData
+from typing import Union, BinaryIO, Optional
+from .data_types import TextData, BinaryData, S3BinaryData
 from tempfile import SpooledTemporaryFile
-
+from s3fs.core import S3File, S3FileSystem
 from cognee.modules.ingestion.exceptions import IngestionError
 
 
-def classify(data: Union[str, BinaryIO], filename: str = None):
+def classify(data: Union[str, BinaryIO], filename: str = None, s3fs: Optional[S3FileSystem] = None):
     if isinstance(data, str):
         return TextData(data)
 
     if isinstance(data, BufferedReader) or isinstance(data, SpooledTemporaryFile):
         return BinaryData(data, str(data.name).split("/")[-1] if data.name else filename)
 
+    if isinstance(data, S3File):
+        derived_filename = str(data.full_name).split("/")[-1] if data.full_name else filename
+        return S3BinaryData(s3_path=data.full_name, name=derived_filename, s3=s3fs)
+
     raise IngestionError(
         message=f"Type of data sent to classify(data: Union[str, BinaryIO) not supported: {type(data)}"
     )
diff --git a/cognee/modules/ingestion/data_types/S3BinaryData.py b/cognee/modules/ingestion/data_types/S3BinaryData.py
new file mode 100644
index 000000000..0536665c8
--- /dev/null
+++ b/cognee/modules/ingestion/data_types/S3BinaryData.py
@@ -0,0 +1,42 @@
+from typing import Optional
+import s3fs
+from cognee.infrastructure.files import get_file_metadata, FileMetadata
+from .IngestionData import IngestionData
+
+
+def create_s3_binary_data(
+    s3_path: str, name: Optional[str] = None, s3: Optional[s3fs.S3FileSystem] = None
+) -> "S3BinaryData":
+    return S3BinaryData(s3_path, name=name, s3=s3)
+
+
+class S3BinaryData(IngestionData):
+    name: Optional[str] = None
+    s3_path: str = None
+    fs: s3fs.S3FileSystem = None
+    metadata: Optional[FileMetadata] = None
+
+    def __init__(
+        self, s3_path: str, name: Optional[str] = None, s3: Optional[s3fs.S3FileSystem] = None
+    ):
+        self.s3_path = s3_path
+        self.name = name
+        self.fs = s3 if s3 is not None else s3fs.S3FileSystem()
+
+    def get_identifier(self):
+        metadata = self.get_metadata()
+        return metadata["content_hash"]
+
+    def get_metadata(self):
+        self.ensure_metadata()
+        return self.metadata
+
+    def ensure_metadata(self):
+        if self.metadata is None:
+            with self.fs.open(self.s3_path, "rb") as f:
+                self.metadata = get_file_metadata(f)
+            if self.metadata.get("name") is None:
+                self.metadata["name"] = self.name or self.s3_path.split("/")[-1]
+
+    def get_data(self):
+        return self.fs.open(self.s3_path, "rb")
diff --git a/cognee/modules/ingestion/data_types/__init__.py b/cognee/modules/ingestion/data_types/__init__.py
index 2cc5796cd..f68306f80 100644
--- a/cognee/modules/ingestion/data_types/__init__.py
+++ b/cognee/modules/ingestion/data_types/__init__.py
@@ -1,3 +1,4 @@
 from .TextData import TextData, create_text_data
 from .BinaryData import BinaryData, create_binary_data
+from .S3BinaryData import S3BinaryData, create_s3_binary_data
 from .IngestionData import IngestionData
diff --git a/cognee/tasks/ingestion/ingest_data.py b/cognee/tasks/ingestion/ingest_data.py
index 78475d106..fe8a81be0 100644
--- a/cognee/tasks/ingestion/ingest_data.py
+++ b/cognee/tasks/ingestion/ingest_data.py
@@ -1,6 +1,7 @@
 from typing import Any, List
 
 import dlt
+import s3fs
 import cognee.modules.ingestion as ingestion
 from cognee.infrastructure.databases.relational import get_relational_engine
 from cognee.modules.data.methods import create_dataset, get_dataset_data, get_datasets_by_name
@@ -13,6 +14,8 @@ from .save_data_item_to_storage import save_data_item_to_storage
 from typing import Union, BinaryIO
 import inspect
 
+from cognee.api.v1.add.config import get_s3_config
+
 
 async def ingest_data(data: Any, dataset_name: str, user: User):
     destination = get_dlt_destination()
@@ -22,6 +25,21 @@ async def ingest_data(data: Any, dataset_name: str, user: User):
         destination=destination,
     )
 
+    s3_config = get_s3_config()
+
+    fs = None
+    if s3_config.aws_access_key_id is not None and s3_config.aws_secret_access_key is not None:
+        fs = s3fs.S3FileSystem(
+            key=s3_config.aws_access_key_id, secret=s3_config.aws_secret_access_key, anon=False
+        )
+
+    def open_data_file(file_path: str):
+        if file_path.startswith("s3://"):
+            return fs.open(file_path, mode="rb")
+        else:
+            local_path = file_path.replace("file://", "")
+            return open(local_path, mode="rb")
+
     def get_external_metadata_dict(data_item: Union[BinaryIO, str, Any]) -> dict[str, Any]:
         if hasattr(data_item, "dict") and inspect.ismethod(getattr(data_item, "dict")):
             return {"metadata": data_item.dict(), "origin": str(type(data_item))}
@@ -31,8 +49,11 @@ async def ingest_data(data: Any, dataset_name: str, user: User):
     @dlt.resource(standalone=True, primary_key="id", merge_key="id")
     async def data_resources(file_paths: List[str], user: User):
         for file_path in file_paths:
-            with open(file_path.replace("file://", ""), mode="rb") as file:
-                classified_data = ingestion.classify(file)
+            with open_data_file(file_path) as file:
+                if file_path.startswith("s3://"):
+                    classified_data = ingestion.classify(file, s3fs=fs)
+                else:
+                    classified_data = ingestion.classify(file)
                 data_id = ingestion.identify(classified_data, user)
                 file_metadata = classified_data.get_metadata()
                 yield {
@@ -59,8 +80,9 @@ async def ingest_data(data: Any, dataset_name: str, user: User):
             file_paths.append(file_path)
 
             # Ingest data and add metadata
-            with open(file_path.replace("file://", ""), mode="rb") as file:
-                classified_data = ingestion.classify(file)
+            # with open(file_path.replace("file://", ""), mode="rb") as file:
+            with open_data_file(file_path) as file:
+                classified_data = ingestion.classify(file, s3fs=fs)
 
                 # data_id is the hash of file contents + owner id to avoid duplicate data
                 data_id = ingestion.identify(classified_data, user)
diff --git a/cognee/tasks/ingestion/resolve_data_directories.py b/cognee/tasks/ingestion/resolve_data_directories.py
index e549688d8..0b3edc382 100644
--- a/cognee/tasks/ingestion/resolve_data_directories.py
+++ b/cognee/tasks/ingestion/resolve_data_directories.py
@@ -1,5 +1,8 @@
 import os
+import s3fs
 from typing import List, Union, BinaryIO
+from urllib.parse import urlparse
+from cognee.api.v1.add.config import get_s3_config
 
 
 async def resolve_data_directories(
@@ -20,10 +23,35 @@ async def resolve_data_directories(
         data = [data]
 
     resolved_data = []
+    s3_config = get_s3_config()
+
+    fs = None
+    if s3_config.aws_access_key_id is not None and s3_config.aws_secret_access_key is not None:
+        fs = s3fs.S3FileSystem(
+            key=s3_config.aws_access_key_id, secret=s3_config.aws_secret_access_key, anon=False
+        )
 
     for item in data:
         if isinstance(item, str):  # Check if the item is a path
-            if os.path.isdir(item):  # If it's a directory
+            # S3
+            if urlparse(item).scheme == "s3":
+                if fs is not None:
+                    if include_subdirectories:
+                        base_path = item if item.endswith("/") else item + "/"
+                        s3_keys = fs.glob(base_path + "**")
+                    else:
+                        s3_keys = fs.ls(item)
+                    # Filter out keys that represent directories using fs.isdir
+                    s3_files = []
+                    for key in s3_keys:
+                        if not fs.isdir(key):
+                            if not key.startswith("s3://"):
+                                s3_files.append("s3://" + key)
+                            else:
+                                s3_files.append(key)
+                    resolved_data.extend(s3_files)
+
+            elif os.path.isdir(item):  # If it's a directory
                 if include_subdirectories:
                     # Recursively add all files in the directory and subdirectories
                     for root, _, files in os.walk(item):
diff --git a/cognee/tasks/ingestion/save_data_item_to_storage.py b/cognee/tasks/ingestion/save_data_item_to_storage.py
index 3f9d572c9..976c1dd26 100644
--- a/cognee/tasks/ingestion/save_data_item_to_storage.py
+++ b/cognee/tasks/ingestion/save_data_item_to_storage.py
@@ -16,8 +16,10 @@ async def save_data_item_to_storage(data_item: Union[BinaryIO, str, Any], datase
         file_path = save_data_to_file(data_item.file, filename=data_item.filename)
 
     elif isinstance(data_item, str):
+        if data_item.startswith("s3://"):
+            file_path = data_item
         # data is a file path
-        if data_item.startswith("file://") or data_item.startswith("/"):
+        elif data_item.startswith("file://") or data_item.startswith("/"):
             file_path = data_item.replace("file://", "")
         # data is text
         else:
diff --git a/cognee/tests/test_s3.py b/cognee/tests/test_s3.py
new file mode 100644
index 000000000..888458d33
--- /dev/null
+++ b/cognee/tests/test_s3.py
@@ -0,0 +1,86 @@
+import logging
+
+import cognee
+from cognee.shared.logging_utils import get_logger
+from cognee.infrastructure.databases.graph.get_graph_engine import get_graph_engine
+from collections import Counter
+
+
+logger = get_logger()
+
+
+async def main():
+    await cognee.prune.prune_data()
+    await cognee.prune.prune_system(metadata=True)
+
+    s3_input = "s3://samples3input"
+    await cognee.add(s3_input)
+
+    await cognee.cognify()
+
+    graph_engine = await get_graph_engine()
+    graph = await graph_engine.get_graph()
+
+    type_counts = Counter(
+        node_data["type"] for _, node_data in graph.nodes(data=True) if "type" in node_data
+    )
+
+    edge_type_counts = Counter(edge_type for _, _, edge_type in graph.edges(keys=True))
+
+    logging.info(type_counts)
+
+    # Assert there is exactly one PdfDocument.
+    assert type_counts.get("PdfDocument", 0) == 1, (
+        f"Expected exactly one PdfDocument, but found {type_counts.get('PdfDocument', 0)}"
+    )
+
+    # Assert there is exactly one TextDocument.
+    assert type_counts.get("TextDocument", 0) == 1, (
+        f"Expected exactly one TextDocument, but found {type_counts.get('TextDocument', 0)}"
+    )
+
+    # Assert there are at least two DocumentChunk nodes.
+    assert type_counts.get("DocumentChunk", 0) >= 2, (
+        f"Expected at least two DocumentChunk nodes, but found {type_counts.get('DocumentChunk', 0)}"
+    )
+
+    # Assert there is at least two TextSummary.
+    assert type_counts.get("TextSummary", 0) >= 2, (
+        f"Expected at least two TextSummary, but found {type_counts.get('TextSummary', 0)}"
+    )
+
+    # Assert there is at least one Entity.
+    assert type_counts.get("Entity", 0) > 0, (
+        f"Expected more than zero Entity nodes, but found {type_counts.get('Entity', 0)}"
+    )
+
+    # Assert there is at least one EntityType.
+    assert type_counts.get("EntityType", 0) > 0, (
+        f"Expected more than zero EntityType nodes, but found {type_counts.get('EntityType', 0)}"
+    )
+
+    # Assert that there are at least two 'is_part_of' edges.
+    assert edge_type_counts.get("is_part_of", 0) >= 2, (
+        f"Expected at least two 'is_part_of' edges, but found {edge_type_counts.get('is_part_of', 0)}"
+    )
+
+    # Assert that there are at least two 'made_from' edges.
+    assert edge_type_counts.get("made_from", 0) >= 2, (
+        f"Expected at least two 'made_from' edges, but found {edge_type_counts.get('made_from', 0)}"
+    )
+
+    # Assert that there is at least one 'is_a' edge.
+    assert edge_type_counts.get("is_a", 0) >= 1, (
+        f"Expected at least one 'is_a' edge, but found {edge_type_counts.get('is_a', 0)}"
+    )
+
+    # Assert that there is at least one 'contains' edge.
+    assert edge_type_counts.get("contains", 0) >= 1, (
+        f"Expected at least one 'contains' edge, but found {edge_type_counts.get('contains', 0)}"
+    )
+
+
+if __name__ == "__main__":
+    import asyncio
+
+    asyncio.run(main())
diff --git a/poetry.lock b/poetry.lock
index 317054fe0..5879918aa 100644
--- a/poetry.lock
+++ b/poetry.lock
@@ -1,4 +1,30 @@
-# This file is automatically @generated by Poetry 2.1.2 and should not be changed by hand.
+# This file is automatically @generated by Poetry 2.1.1 and should not be changed by hand.
+
+[[package]]
+name = "aiobotocore"
+version = "2.21.1"
+description = "Async client for aws services using botocore and aiohttp"
+optional = false
+python-versions = ">=3.8"
+groups = ["main"]
+markers = "extra == \"filesystem\""
+files = [
+    {file = "aiobotocore-2.21.1-py3-none-any.whl", hash = "sha256:bd7c49a6d6f8a3d9444b0a94417c8da13813b5c7eec1c4f0ec2db7e8ce8f23e7"},
+    {file = "aiobotocore-2.21.1.tar.gz", hash = "sha256:010357f43004413e92a9d066bb0db1f241aeb29ffed306e9197061ffc94e6577"},
+]
+
+[package.dependencies]
+aiohttp = ">=3.9.2,<4.0.0"
+aioitertools = ">=0.5.1,<1.0.0"
+botocore = ">=1.37.0,<1.37.2"
+jmespath = ">=0.7.1,<2.0.0"
+multidict = ">=6.0.0,<7.0.0"
+python-dateutil = ">=2.1,<3.0.0"
+wrapt = ">=1.10.10,<2.0.0"
+
+[package.extras]
+awscli = ["awscli (>=1.38.0,<1.38.2)"]
+boto3 = ["boto3 (>=1.37.0,<1.37.2)"]
 
 [[package]]
 name = "aiofiles"
@@ -128,6 +154,23 @@ yarl = ">=1.17.0,<2.0"
 [package.extras]
 speedups = ["Brotli ; platform_python_implementation == \"CPython\"", "aiodns (>=3.2.0) ; sys_platform == \"linux\" or sys_platform == \"darwin\"", "brotlicffi ; platform_python_implementation != \"CPython\""]
 
+[[package]]
+name = "aioitertools"
+version = "0.12.0"
+description = "itertools and builtins for AsyncIO and mixed iterables"
+optional = false
+python-versions = ">=3.8"
+groups = ["main"]
+markers = "extra == \"filesystem\""
+files = [
+    {file = "aioitertools-0.12.0-py3-none-any.whl", hash = "sha256:fc1f5fac3d737354de8831cbba3eb04f79dd649d8f3afb4c5b114925e662a796"},
+    {file = "aioitertools-0.12.0.tar.gz", hash = "sha256:c2a9055b4fbb7705f561b9d86053e8af5d10cc845d22c32008c43490b2d8dd6b"},
+]
+
+[package.extras]
+dev = ["attribution (==1.8.0)", "black (==24.8.0)", "build (>=1.2)", "coverage (==7.6.1)", "flake8 (==7.1.1)", "flit (==3.9.0)", "mypy (==1.11.2)", "ufmt (==2.7.1)", "usort (==1.0.8.post1)"]
+docs = ["sphinx (==8.0.2)", "sphinx-mdinclude (==0.6.2)"]
+
 [[package]]
 name = "aiosignal"
 version = "1.3.2"
@@ -408,7 +451,7 @@ description = "Timeout context manager for asyncio programs"
 optional = false
 python-versions = ">=3.7"
 groups = ["main"]
-markers = "python_version == \"3.10\""
+markers = "python_version < \"3.11\""
 files = [
     {file = "async-timeout-4.0.3.tar.gz", hash = "sha256:4640d96be84d82d02ed59ea2b7105a0f7b33abe8703703cd0ab0bf87c427522f"},
     {file = "async_timeout-4.0.3-py3-none-any.whl", hash = "sha256:7405140ff1230c310e51dc27b3145b9092d659ce68ff733fb0cefe3ee42be028"},
@@ -565,7 +608,7 @@ description = "Backport of CPython tarfile module"
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "python_version < \"3.12\" and extra == \"deepeval\""
+markers = "extra == \"deepeval\" and python_version <= \"3.11\""
 files = [
     {file = "backports.tarfile-1.2.0-py3-none-any.whl", hash = "sha256:77e284d754527b01fb1e6fa8a1afe577858ebe4e9dad8919e34c862cb399bc34"},
     {file = "backports_tarfile-1.2.0.tar.gz", hash = "sha256:d75e02c268746e1b8144c278978b6e98e85de6ad16f8e4b0844a154557eca991"},
@@ -774,18 +817,18 @@ css = ["tinycss2 (>=1.1.0,<1.5)"]
 
 [[package]]
 name = "boto3"
-version = "1.37.33"
+version = "1.37.1"
 description = "The AWS SDK for Python"
 optional = false
 python-versions = ">=3.8"
 groups = ["main"]
 files = [
-    {file = "boto3-1.37.33-py3-none-any.whl", hash = "sha256:7b1b1bc69762975824e5a5d570880abebf634f7594f88b3dc175e8800f35be1a"},
-    {file = "boto3-1.37.33.tar.gz", hash = "sha256:4390317a1578af73f1514651bd180ba25802dcbe0a23deafa13851d54d3c3203"},
+    {file = "boto3-1.37.1-py3-none-any.whl", hash = "sha256:4320441f904435a1b85e6ecb81793192e522c737cc9ed6566014e29f0a11cb22"},
+    {file = "boto3-1.37.1.tar.gz", hash = "sha256:96d18f7feb0c1fcb95f8837b74b6c8880e1b4e35ce5f8a8f8cb243a090c278ed"},
 ]
 
 [package.dependencies]
-botocore = ">=1.37.33,<1.38.0"
+botocore = ">=1.37.1,<1.38.0"
 jmespath = ">=0.7.1,<2.0.0"
 s3transfer = ">=0.11.0,<0.12.0"
 
@@ -794,14 +837,14 @@ crt = ["botocore[crt] (>=1.21.0,<2.0a0)"]
 
 [[package]]
 name = "botocore"
-version = "1.37.33"
+version = "1.37.1"
 description = "Low-level, data-driven core of boto 3."
 optional = false
 python-versions = ">=3.8"
 groups = ["main"]
 files = [
-    {file = "botocore-1.37.33-py3-none-any.whl", hash = "sha256:4a167dfecae51e9140de24067de1c339acde5ade3dad524a4600ac2c72055e23"},
-    {file = "botocore-1.37.33.tar.gz", hash = "sha256:09b213b0d0500040f85c7daee912ea767c724e43ed61909e624c803ff6925222"},
+    {file = "botocore-1.37.1-py3-none-any.whl", hash = "sha256:c1db1bfc5d8c6b3b6d1ca6794f605294b4264e82a7e727b88e0fef9c2b9fbb9c"},
+    {file = "botocore-1.37.1.tar.gz", hash = "sha256:b194db8fb2a0ffba53568c364ae26166e7eec0445496b2ac86a6e142f3dd982f"},
 ]
 
 [package.dependencies]
@@ -1183,7 +1226,7 @@ files = [
     {file = "colorama-0.4.6-py2.py3-none-any.whl", hash = "sha256:4f1d9991f5acc0ca119f9d443620b77f9d6b33703e51011c16baf57afb285fc6"},
     {file = "colorama-0.4.6.tar.gz", hash = "sha256:08695f5cb7ed6e0531a20572697297273c47b8cae5a63ffc6d6ed5c201be6e44"},
 ]
-markers = {main = "(platform_system == \"Windows\" or sys_platform == \"win32\" or os_name == \"nt\" or extra == \"llama-index\" or extra == \"deepeval\") and (platform_system == \"Windows\" or sys_platform == \"win32\" or extra == \"chromadb\" or extra == \"llama-index\" or extra == \"deepeval\")", dev = "platform_system == \"Windows\" or sys_platform == \"win32\""}
+markers = {main = "(platform_system == \"Windows\" or sys_platform == \"win32\" or extra == \"chromadb\" or extra == \"llama-index\" or extra == \"deepeval\") and (os_name == \"nt\" or platform_system == \"Windows\" or sys_platform == \"win32\" or extra == \"llama-index\" or extra == \"deepeval\")", dev = "platform_system == \"Windows\" or sys_platform == \"win32\""}
 
 [[package]]
 name = "coloredlogs"
@@ -1192,7 +1235,7 @@ description = "Colored terminal output for Python's logging module"
 optional = true
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "coloredlogs-15.0.1-py2.py3-none-any.whl", hash = "sha256:612ee75c546f53e92e70049c9dbfcc18c935a2b9a53b66085ce9ef6a6e5c0934"},
     {file = "coloredlogs-15.0.1.tar.gz", hash = "sha256:7c991aa71a4577af2f82600d8f8f3a89f936baeaf9b50a9c197da014e5bf16b0"},
@@ -1806,7 +1849,7 @@ description = "Python datetimes made easy"
 optional = false
 python-versions = ">=3.9"
 groups = ["main"]
-markers = "python_full_version == \"3.13.0\""
+markers = "python_version >= \"3.13\""
 files = [
     {file = "dlt_pendulum-3.0.2-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:9ae1222828474f9e4743f8929f8026abe2d0b3a99427a483da2868690b017332"},
     {file = "dlt_pendulum-3.0.2-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:75e1b758f88f887706902438fa5b293f11cec5d656c6540c9957da8c9b953198"},
@@ -1989,7 +2032,7 @@ description = "Backport of PEP 654 (exception groups)"
 optional = false
 python-versions = ">=3.7"
 groups = ["main", "dev"]
-markers = "python_version == \"3.10\""
+markers = "python_version < \"3.11\""
 files = [
     {file = "exceptiongroup-1.2.2-py3-none-any.whl", hash = "sha256:3111b9d131c238bec2f8f516e123e14ba243563fb135d3fe885990585aa7795b"},
     {file = "exceptiongroup-1.2.2.tar.gz", hash = "sha256:47c2edf7c6738fafb49fd34290706d1a1a2f4d1c6df275526b62cbb4aa5393cc"},
@@ -2115,7 +2158,7 @@ description = "Fast, light, accurate library built for retrieval embedding gener
 optional = true
 python-versions = ">=3.9.0"
 groups = ["main"]
-markers = "extra == \"codegraph\" and python_version < \"3.13\""
+markers = "python_version < \"3.13\" and extra == \"codegraph\""
 files = [
     {file = "fastembed-0.6.0-py3-none-any.whl", hash = "sha256:a08385e9388adea0529a586004f2d588c9787880a510e4e5d167127a11e75328"},
     {file = "fastembed-0.6.0.tar.gz", hash = "sha256:5c9ead25f23449535b07243bbe1f370b820dcc77ec2931e61674e3fe7ff24733"},
@@ -2187,7 +2230,7 @@ description = "The FlatBuffers serialization format for Python"
 optional = true
 python-versions = "*"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "flatbuffers-25.2.10-py2.py3-none-any.whl", hash = "sha256:ebba5f4d5ea615af3f7fd70fc310636fbb2bbd1f566ac0a23d98dd412de50051"},
     {file = "flatbuffers-25.2.10.tar.gz", hash = "sha256:97e451377a41262f8d9bd4295cc836133415cc03d8cb966410a4af92eb00d26e"},
@@ -2661,7 +2704,7 @@ description = "Common protobufs used in Google APIs"
 optional = true
 python-versions = ">=3.7"
 groups = ["main"]
-markers = "extra == \"gemini\" or extra == \"chromadb\" or extra == \"deepeval\""
+markers = "extra == \"gemini\" or extra == \"chromadb\" or extra == \"deepeval\" or python_version < \"3.11\" and (extra == \"chromadb\" or extra == \"deepeval\" or extra == \"gemini\")"
 files = [
     {file = "googleapis_common_protos-1.70.0-py3-none-any.whl", hash = "sha256:b8bfcca8c25a2bb253e0e0b0adaf8c00773e5e6af6fd92397576680b807e0fd8"},
     {file = "googleapis_common_protos-1.70.0.tar.gz", hash = "sha256:0e1b44e0ea153e6594f9f394fef15193a68aaaea2d843f83e2742717ca753257"},
@@ -2866,7 +2909,7 @@ description = "HTTP/2-based RPC framework"
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "extra == \"gemini\" or extra == \"chromadb\" or extra == \"deepeval\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"milvus\""
+markers = "extra == \"gemini\" or extra == \"chromadb\" or extra == \"deepeval\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"milvus\" or python_version < \"3.11\" and (extra == \"chromadb\" or extra == \"deepeval\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"milvus\" or extra == \"gemini\")"
 files = [
     {file = "grpcio-1.67.1-cp310-cp310-linux_armv7l.whl", hash = "sha256:8b0341d66a57f8a3119b77ab32207072be60c9bf79760fa609c5609f2deb1f3f"},
     {file = "grpcio-1.67.1-cp310-cp310-macosx_12_0_universal2.whl", hash = "sha256:f5a27dddefe0e2357d3e617b9079b4bfdc91341a91565111a21ed6ebbc51b22d"},
@@ -3326,7 +3369,7 @@ description = "Human friendly output for text interfaces using Python"
 optional = true
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "humanfriendly-10.0-py2.py3-none-any.whl", hash = "sha256:1697e1a8a8f550fd43c2865cd84542fc175a61dcb779b6fee18cf6b6ccba1477"},
     {file = "humanfriendly-10.0.tar.gz", hash = "sha256:6b0b831ce8f15f7300721aa49829fc4e83921a9a301cc7f606be6686a2288ddc"},
@@ -3527,7 +3570,7 @@ description = "IPython: Productive Interactive Computing"
 optional = true
 python-versions = ">=3.10"
 groups = ["main", "dev"]
-markers = "python_version == \"3.10\""
+markers = "python_version < \"3.11\""
 files = [
     {file = "ipython-8.35.0-py3-none-any.whl", hash = "sha256:e6b7470468ba6f1f0a7b116bb688a3ece2f13e2f94138e508201fad677a788ba"},
     {file = "ipython-8.35.0.tar.gz", hash = "sha256:d200b7d93c3f5883fc36ab9ce28a18249c7706e51347681f80a0aef9895f2520"},
@@ -3732,7 +3775,7 @@ description = "Low-level, pure Python DBus protocol wrapper."
 optional = true
 python-versions = ">=3.7"
 groups = ["main"]
-markers = "sys_platform == \"linux\" and extra == \"deepeval\""
+markers = "extra == \"deepeval\" and sys_platform == \"linux\""
 files = [
     {file = "jeepney-0.9.0-py3-none-any.whl", hash = "sha256:97e5714520c16fc0a45695e5365a2e11b81ea79bba796e26f9f1d178cb182683"},
     {file = "jeepney-0.9.0.tar.gz", hash = "sha256:cf0e9e845622b81e4a28df94c40345400256ec608d0e55bb8a3feaa9163f5732"},
@@ -4999,7 +5042,7 @@ description = "Python logging made (stupidly) simple"
 optional = true
 python-versions = "<4.0,>=3.5"
 groups = ["main"]
-markers = "extra == \"codegraph\" and python_version < \"3.13\""
+markers = "python_version < \"3.13\" and extra == \"codegraph\""
 files = [
     {file = "loguru-0.7.3-py3-none-any.whl", hash = "sha256:31a33c10c8e1e10422bfd431aeb5d351c7cf7fa671e3c4df004162264b28220c"},
     {file = "loguru-0.7.3.tar.gz", hash = "sha256:19480589e77d47b8d85b2c827ad95d49bf31b0dcde16593892eb51dd18706eb6"},
@@ -5450,7 +5493,7 @@ description = "A lightweight version of Milvus wrapped with Python."
 optional = true
 python-versions = ">=3.7"
 groups = ["main"]
-markers = "sys_platform != \"win32\" and extra == \"milvus\""
+markers = "extra == \"milvus\" and sys_platform != \"win32\""
 files = [
     {file = "milvus_lite-2.4.12-py3-none-macosx_10_9_x86_64.whl", hash = "sha256:e8d4f7cdd5f731efd6faeee3715d280fd91a5f9b4d89312664d56401f65b1473"},
     {file = "milvus_lite-2.4.12-py3-none-macosx_11_0_arm64.whl", hash = "sha256:20087663e7b4385050b7ad08f1f03404426d4c87b1ff91d5a8723eee7fd49e88"},
@@ -5805,7 +5848,7 @@ description = "Python library for arbitrary-precision floating-point arithmetic"
 optional = true
 python-versions = "*"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "mpmath-1.3.0-py3-none-any.whl", hash = "sha256:a0b2b9fe80bbcd81a6647ff13108738cfb482d481d826cc0e02f5b35e5c88d2c"},
     {file = "mpmath-1.3.0.tar.gz", hash = "sha256:7a28eb2a9774d00c7bc92411c19a89209d5da7c4c9a9e227be8330a23a25b91f"},
@@ -6380,7 +6423,7 @@ description = "ONNX Runtime is a runtime accelerator for Machine Learning models
 optional = true
 python-versions = ">=3.10"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "onnxruntime-1.21.0-cp310-cp310-macosx_13_0_universal2.whl", hash = "sha256:95513c9302bc8dd013d84148dcf3168e782a80cdbf1654eddc948a23147ccd3d"},
     {file = "onnxruntime-1.21.0-cp310-cp310-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:635d4ab13ae0f150dd4c6ff8206fd58f1c6600636ecc796f6f0c42e4c918585b"},
@@ -6853,8 +6896,8 @@ files = [
 [package.dependencies]
 numpy = [
     {version = ">=1.22.4", markers = "python_version < \"3.11\""},
-    {version = ">=1.23.2", markers = "python_version == \"3.11\""},
     {version = ">=1.26.0", markers = "python_version >= \"3.12\""},
+    {version = ">=1.23.2", markers = "python_version == \"3.11\""},
 ]
 python-dateutil = ">=2.8.2"
 pytz = ">=2020.1"
@@ -7488,7 +7531,7 @@ description = ""
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"gemini\" or extra == \"deepeval\" or extra == \"milvus\") and python_version < \"3.13\" or extra == \"gemini\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"chromadb\" or extra == \"deepeval\" or extra == \"milvus\""
+markers = "python_version == \"3.10\" and extra == \"codegraph\" or (extra == \"chromadb\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"deepeval\" or extra == \"gemini\" or extra == \"milvus\") and python_version < \"3.11\" or (python_version == \"3.12\" or extra == \"gemini\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"chromadb\" or extra == \"deepeval\" or extra == \"milvus\") and (extra == \"codegraph\" or extra == \"gemini\" or extra == \"chromadb\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"deepeval\" or extra == \"milvus\") and python_version >= \"3.12\" or python_version == \"3.11\" and (extra == \"codegraph\" or extra == \"gemini\" or extra == \"chromadb\" or extra == \"qdrant\" or extra == \"weaviate\" or extra == \"deepeval\" or extra == \"milvus\")"
 files = [
     {file = "protobuf-5.29.4-cp310-abi3-win32.whl", hash = "sha256:13eb236f8eb9ec34e63fc8b1d6efd2777d062fa6aaa68268fb67cf77f6839ad7"},
     {file = "protobuf-5.29.4-cp310-abi3-win_amd64.whl", hash = "sha256:bcefcdf3976233f8a502d265eb65ea740c989bacc6c30a58290ed0e519eb4b8d"},
@@ -7602,7 +7645,7 @@ description = "Fast and parallel snowball stemmer"
 optional = true
 python-versions = "*"
 groups = ["main"]
-markers = "extra == \"codegraph\" and python_version < \"3.13\""
+markers = "python_version < \"3.13\" and extra == \"codegraph\""
 files = [
     {file = "py_rust_stemmers-0.1.5-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:bfbd9034ae00419ff2154e33b8f5b4c4d99d1f9271f31ed059e5c7e9fa005844"},
     {file = "py_rust_stemmers-0.1.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c7162ae66df2bb0fc39b350c24a049f5f5151c03c046092ba095c2141ec223a2"},
@@ -8004,8 +8047,8 @@ astroid = ">=3.3.8,<=3.4.0.dev0"
 colorama = {version = ">=0.4.5", markers = "sys_platform == \"win32\""}
 dill = [
     {version = ">=0.2", markers = "python_version < \"3.11\""},
-    {version = ">=0.3.6", markers = "python_version >= \"3.11\""},
     {version = ">=0.3.7", markers = "python_version >= \"3.12\""},
+    {version = ">=0.3.6", markers = "python_version == \"3.11\""},
 ]
 isort = ">=4.2.5,<5.13 || >5.13,<7"
 mccabe = ">=0.6,<0.8"
@@ -8146,7 +8189,7 @@ description = "A python implementation of GNU readline."
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "sys_platform == \"win32\" and (extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or sys_platform == \"win32\" and extra == \"chromadb\""
+markers = "sys_platform == \"win32\" and (python_version == \"3.10\" or extra == \"chromadb\" or extra == \"codegraph\") and (extra == \"chromadb\" or python_version == \"3.12\" or python_version == \"3.10\" or python_version == \"3.11\") and (extra == \"codegraph\" or extra == \"chromadb\")"
 files = [
     {file = "pyreadline3-3.5.4-py3-none-any.whl", hash = "sha256:eaf8e6cc3c49bcccf145fc6067ba8643d1df34d604a1ec0eccbf7a18e6d3fae6"},
     {file = "pyreadline3-3.5.4.tar.gz", hash = "sha256:8d57d53039a1c75adba8e50dd3d992b28143480816187ea5efbd5c78e6c885b7"},
@@ -8503,7 +8546,7 @@ files = [
     {file = "pywin32-310-cp39-cp39-win32.whl", hash = "sha256:851c8d927af0d879221e616ae1f66145253537bbdd321a77e8ef701b443a9a1a"},
     {file = "pywin32-310-cp39-cp39-win_amd64.whl", hash = "sha256:96867217335559ac619f00ad70e513c0fcf84b8a3af9fc2bba3b59b97da70475"},
 ]
-markers = {main = "(platform_system == \"Windows\" or sys_platform == \"win32\") and (extra == \"qdrant\" or extra == \"deepeval\" or sys_platform == \"win32\")", dev = "platform_python_implementation != \"PyPy\" and sys_platform == \"win32\""}
+markers = {main = "(extra == \"qdrant\" or extra == \"deepeval\") and platform_system == \"Windows\" or sys_platform == \"win32\"", dev = "sys_platform == \"win32\" and platform_python_implementation != \"PyPy\""}
 
 [[package]]
 name = "pywin32-ctypes"
@@ -8512,7 +8555,7 @@ description = "A (partial) reimplementation of pywin32 using ctypes/cffi"
 optional = true
 python-versions = ">=3.6"
 groups = ["main"]
-markers = "sys_platform == \"win32\" and extra == \"deepeval\""
+markers = "extra == \"deepeval\" and sys_platform == \"win32\""
 files = [
     {file = "pywin32-ctypes-0.2.3.tar.gz", hash = "sha256:d162dc04946d704503b2edc4d55f3dba5c1d539ead017afa00142c38b9885755"},
     {file = "pywin32_ctypes-0.2.3-py3-none-any.whl", hash = "sha256:8a1513379d709975552d202d942d9837758905c8d01eb82b8bcc30918929e7b8"},
@@ -8740,7 +8783,7 @@ description = "Client library for the Qdrant vector search engine"
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "python_full_version == \"3.13.0\" and extra == \"qdrant\""
+markers = "python_version >= \"3.13\" and extra == \"qdrant\""
 files = [
     {file = "qdrant_client-1.12.1-py3-none-any.whl", hash = "sha256:b2d17ce18e9e767471368380dd3bbc4a0e3a0e2061fedc9af3542084b48451e0"},
     {file = "qdrant_client-1.12.1.tar.gz", hash = "sha256:35e8e646f75b7b883b3d2d0ee4c69c5301000bba41c82aa546e985db0f1aeb72"},
@@ -8766,7 +8809,7 @@ description = "Client library for the Qdrant vector search engine"
 optional = true
 python-versions = ">=3.9"
 groups = ["main"]
-markers = "extra == \"qdrant\" and python_version < \"3.13\""
+markers = "python_version < \"3.13\" and extra == \"qdrant\""
 files = [
     {file = "qdrant_client-1.13.3-py3-none-any.whl", hash = "sha256:f52cacbb936e547d3fceb1aaed3e3c56be0ebfd48e8ea495ea3dbc89c671d1d2"},
     {file = "qdrant_client-1.13.3.tar.gz", hash = "sha256:61ca09e07c6d7ac0dfbdeb13dca4fe5f3e08fa430cb0d74d66ef5d023a70adfc"},
@@ -9379,23 +9422,45 @@ files = [
     {file = "ruff-0.11.5.tar.gz", hash = "sha256:cae2e2439cb88853e421901ec040a758960b576126dab520fa08e9de431d1bef"},
 ]
 
+[[package]]
+name = "s3fs"
+version = "2025.3.2"
+description = "Convenient Filesystem interface over S3"
+optional = false
+python-versions = ">=3.9"
+groups = ["main"]
+markers = "extra == \"filesystem\""
+files = [
+    {file = "s3fs-2025.3.2-py3-none-any.whl", hash = "sha256:81eae3f37b4b04bcc08845d7bcc607c6ca45878813ef7e6a28d77b2688417130"},
+    {file = "s3fs-2025.3.2.tar.gz", hash = "sha256:6798f896ec76dd3bfd8beb89f0bb7c5263cb2760e038bae0978505cd172a307c"},
+]
+
+[package.dependencies]
+aiobotocore = ">=2.5.4,<3.0.0"
+aiohttp = "<4.0.0a0 || >4.0.0a0,<4.0.0a1 || >4.0.0a1"
+fsspec = "==2025.3.2.*"
+
+[package.extras]
+awscli = ["aiobotocore[awscli] (>=2.5.4,<3.0.0)"]
+boto3 = ["aiobotocore[boto3] (>=2.5.4,<3.0.0)"]
+
 [[package]]
 name = "s3transfer"
-version = "0.11.4"
+version = "0.11.3"
 description = "An Amazon S3 Transfer Manager"
 optional = false
 python-versions = ">=3.8"
 groups = ["main"]
 files = [
-    {file = "s3transfer-0.11.4-py3-none-any.whl", hash = "sha256:ac265fa68318763a03bf2dc4f39d5cbd6a9e178d81cc9483ad27da33637e320d"},
-    {file = "s3transfer-0.11.4.tar.gz", hash = "sha256:559f161658e1cf0a911f45940552c696735f5c74e64362e515f333ebed87d679"},
+    {file = "s3transfer-0.11.3-py3-none-any.whl", hash = "sha256:ca855bdeb885174b5ffa95b9913622459d4ad8e331fc98eb01e6d5eb6a30655d"},
+    {file = "s3transfer-0.11.3.tar.gz", hash = "sha256:edae4977e3a122445660c7c114bba949f9d191bae3b34a096f18a1c8c354527a"},
 ]
 
 [package.dependencies]
-botocore = ">=1.37.4,<2.0a.0"
+botocore = ">=1.36.0,<2.0a.0"
 
 [package.extras]
-crt = ["botocore[crt] (>=1.37.4,<2.0a.0)"]
+crt = ["botocore[crt] (>=1.36.0,<2.0a.0)"]
 
 [[package]]
 name = "safetensors"
@@ -9562,7 +9627,7 @@ description = "Python bindings to FreeDesktop.org Secret Service API"
 optional = true
 python-versions = ">=3.6"
 groups = ["main"]
-markers = "sys_platform == \"linux\" and extra == \"deepeval\""
+markers = "extra == \"deepeval\" and sys_platform == \"linux\""
 files = [
     {file = "SecretStorage-3.3.3-py3-none-any.whl", hash = "sha256:f356e6628222568e3af06f2eba8df495efa13b3b63081dafd4f7d9a7b7bc9f99"},
     {file = "SecretStorage-3.3.3.tar.gz", hash = "sha256:2403533ef369eca6d2ba81718576c5e0f564d5cca1b58f73a8b23e7d4eeebd77"},
@@ -10143,7 +10208,7 @@ description = "Computer algebra system (CAS) in Python"
 optional = true
 python-versions = ">=3.8"
 groups = ["main"]
-markers = "(extra == \"codegraph\" or extra == \"chromadb\") and python_version < \"3.13\" or extra == \"chromadb\""
+markers = "python_version == \"3.10\" and (extra == \"chromadb\" or extra == \"codegraph\") or extra == \"chromadb\" or python_version == \"3.12\" and (extra == \"chromadb\" or extra == \"codegraph\") or python_version == \"3.11\" and (extra == \"chromadb\" or extra == \"codegraph\")"
 files = [
     {file = "sympy-1.13.3-py3-none-any.whl", hash = "sha256:54612cf55a62755ee71824ce692986f23c88ffa77207b30c1368eda4a7060f73"},
     {file = "sympy-1.13.3.tar.gz", hash = "sha256:b27fd2c6530e0ab39e275fc9b683895367e51d5da91baa8d3d64db2565fec4d9"},
@@ -10328,7 +10393,7 @@ description = "A lil' TOML parser"
 optional = false
 python-versions = ">=3.8"
 groups = ["main", "dev"]
-markers = "python_version == \"3.10\""
+markers = "python_version < \"3.11\""
 files = [
     {file = "tomli-2.2.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:678e4fa69e4575eb77d103de3df8a895e1591b48e740211bd1067378c69e8249"},
     {file = "tomli-2.2.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:023aa114dd824ade0100497eb2318602af309e5a55595f76b626d6d9f3b7b0a6"},
@@ -10982,7 +11047,7 @@ description = "Fast implementation of asyncio event loop on top of libuv"
 optional = true
 python-versions = ">=3.8.0"
 groups = ["main"]
-markers = "platform_python_implementation != \"PyPy\" and extra == \"chromadb\" and sys_platform != \"win32\" and sys_platform != \"cygwin\""
+markers = "sys_platform != \"win32\" and sys_platform != \"cygwin\" and platform_python_implementation != \"PyPy\" and extra == \"chromadb\""
 files = [
     {file = "uvloop-0.21.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:ec7e6b09a6fdded42403182ab6b832b71f4edaf7f37a9a0e371a01db5f0cb45f"},
     {file = "uvloop-0.21.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:196274f2adb9689a289ad7d65700d37df0c0930fd8e4e743fa4834e850d7719d"},
@@ -11396,7 +11461,7 @@ description = "A small Python utility to set file creation time on Windows"
 optional = true
 python-versions = ">=3.5"
 groups = ["main"]
-markers = "sys_platform == \"win32\" and extra == \"codegraph\" and python_version < \"3.13\""
+markers = "extra == \"codegraph\" and sys_platform == \"win32\" and python_version < \"3.13\""
 files = [
     {file = "win32_setctime-1.2.0-py3-none-any.whl", hash = "sha256:95d644c4e708aba81dc3704a116d8cbc974d70b3bdb8be1d150e36be6e9d1390"},
     {file = "win32_setctime-1.2.0.tar.gz", hash = "sha256:ae1fdf948f5640aae05c511ade119313fb6a30d7eabe25fef9764dca5873c4c0"},
@@ -11656,7 +11721,7 @@ deepeval = ["deepeval"]
 docs = ["unstructured"]
 evals = ["gdown", "plotly"]
 falkordb = ["falkordb"]
-filesystem = ["botocore"]
+filesystem = ["botocore", "s3fs"]
 gemini = ["google-generativeai"]
 graphiti = ["graphiti-core"]
 groq = ["groq"]
@@ -11678,4 +11743,4 @@ weaviate = ["weaviate-client"]
 [metadata]
 lock-version = "2.1"
 python-versions = ">=3.10,<=3.13"
-content-hash = "c46737c9aa42018f91e663459a98c8c10a964dff47b91679bab5085e8b98faa6"
+content-hash = "cef734016cd8fc4430277b30c639d0699361f0553c00fdd8fa0fbadbab0b7e7b"
diff --git a/pyproject.toml b/pyproject.toml
index 6ca55a4c6..99a081e46 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -86,6 +86,7 @@ structlog = "^25.2.0"
 pyside6 = {version = "^6.8.3", optional = true}
 google-generativeai = {version = "^0.8.4", optional = true}
 notebook = {version = "^7.1.0", optional = true}
+s3fs = "^2025.3.2"
 
 
 [tool.poetry.extras]