Feat: add gmail connector (#11549)

### What problem does this PR solve?

_Briefly describe what this PR aims to solve. Include background context
that will help reviewers understand the purpose of the PR._

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

api/apps/connector_app.py

@@ -28,8 +28,8 @@ from api.db import InputType
 from api.db.services.connector_service import ConnectorService, SyncLogsService
 from api.utils.api_utils import get_data_error_result, get_json_result, validate_request
 from common.constants import RetCode, TaskStatus
-from common.data_source.config import GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI, DocumentSource
+from common.data_source.config import GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI, GMAIL_WEB_OAUTH_REDIRECT_URI, DocumentSource
-from common.data_source.google_util.constant import GOOGLE_DRIVE_WEB_OAUTH_POPUP_TEMPLATE, GOOGLE_SCOPES
+from common.data_source.google_util.constant import GOOGLE_WEB_OAUTH_POPUP_TEMPLATE, GOOGLE_SCOPES
 from common.misc_utils import get_uuid
 from rag.utils.redis_conn import REDIS_CONN
 from api.apps import login_required, current_user
@@ -122,12 +122,30 @@ GOOGLE_WEB_FLOW_RESULT_PREFIX = "google_drive_web_flow_result"
 WEB_FLOW_TTL_SECS = 15 * 60
 
 
-def _web_state_cache_key(flow_id: str) -> str:
+def _web_state_cache_key(flow_id: str, source_type: str | None = None) -> str:
-    return f"{GOOGLE_WEB_FLOW_STATE_PREFIX}:{flow_id}"
+    """Return Redis key for web OAuth state.
+
+    The default prefix keeps backward compatibility for Google Drive.
+    When source_type == "gmail", a different prefix is used so that
+    Drive/Gmail flows don't clash in Redis.
+    """
+    if source_type == "gmail":
+        prefix = "gmail_web_flow_state"
+    else:
+        prefix = GOOGLE_WEB_FLOW_STATE_PREFIX
+    return f"{prefix}:{flow_id}"
 
 
-def _web_result_cache_key(flow_id: str) -> str:
+def _web_result_cache_key(flow_id: str, source_type: str | None = None) -> str:
-    return f"{GOOGLE_WEB_FLOW_RESULT_PREFIX}:{flow_id}"
+    """Return Redis key for web OAuth result.
+
+    Mirrors _web_state_cache_key logic for result storage.
+    """
+    if source_type == "gmail":
+        prefix = "gmail_web_flow_result"
+    else:
+        prefix = GOOGLE_WEB_FLOW_RESULT_PREFIX
+    return f"{prefix}:{flow_id}"
 
 
 def _load_credentials(payload: str | dict[str, Any]) -> dict[str, Any]:
@@ -146,19 +164,22 @@ def _get_web_client_config(credentials: dict[str, Any]) -> dict[str, Any]:
     return {"web": web_section}
 
 
-async def _render_web_oauth_popup(flow_id: str, success: bool, message: str):
+async def _render_web_oauth_popup(flow_id: str, success: bool, message: str, source="drive"):
     status = "success" if success else "error"
     auto_close = "window.close();" if success else ""
     escaped_message = escape(message)
     payload_json = json.dumps(
         {
-            "type": "ragflow-google-drive-oauth",
+            # TODO(google-oauth): include connector type (drive/gmail) in payload type if needed
+            "type": f"ragflow-google-{source}-oauth",
             "status": status,
             "flowId": flow_id or "",
             "message": message,
         }
     )
-    html = GOOGLE_DRIVE_WEB_OAUTH_POPUP_TEMPLATE.format(
+    # TODO(google-oauth): title/heading/message may need to reflect drive/gmail based on cached type
+    html = GOOGLE_WEB_OAUTH_POPUP_TEMPLATE.format(
+        title=f"Google {source.capitalize()} Authorization",
         heading="Authorization complete" if success else "Authorization failed",
         message=escaped_message,
         payload_json=payload_json,
@@ -169,20 +190,33 @@ async def _render_web_oauth_popup(flow_id: str, success: bool, message: str):
     return response
 
 
-@manager.route("/google-drive/oauth/web/start", methods=["POST"])  # noqa: F821
+@manager.route("/google/oauth/web/start", methods=["POST"])  # noqa: F821
 @login_required
 @validate_request("credentials")
-async def start_google_drive_web_oauth():
+async def start_google_web_oauth():
-    if not GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI:
+    source = request.args.get("type", "google-drive")
+    if source not in ("google-drive", "gmail"):
+        return get_json_result(code=RetCode.ARGUMENT_ERROR, message="Invalid Google OAuth type.")
+
+    if source == "gmail":
+        redirect_uri = GMAIL_WEB_OAUTH_REDIRECT_URI
+        scopes = GOOGLE_SCOPES[DocumentSource.GMAIL]
+    else:
+        redirect_uri = GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI if source == "google-drive" else GMAIL_WEB_OAUTH_REDIRECT_URI
+        scopes = GOOGLE_SCOPES[DocumentSource.GOOGLE_DRIVE if source == "google-drive" else DocumentSource.GMAIL]
+
+    if not redirect_uri:
         return get_json_result(
             code=RetCode.SERVER_ERROR,
-            message="Google Drive OAuth redirect URI is not configured on the server.",
+            message="Google OAuth redirect URI is not configured on the server.",
         )
 
     req = await request.json or {}
     raw_credentials = req.get("credentials", "")
+
     try:
         credentials = _load_credentials(raw_credentials)
+        print(credentials)
     except ValueError as exc:
         return get_json_result(code=RetCode.ARGUMENT_ERROR, message=str(exc))
 
@@ -199,8 +233,8 @@ async def start_google_drive_web_oauth():
 
     flow_id = str(uuid.uuid4())
     try:
-        flow = Flow.from_client_config(client_config, scopes=GOOGLE_SCOPES[DocumentSource.GOOGLE_DRIVE])
+        flow = Flow.from_client_config(client_config, scopes=scopes)
-        flow.redirect_uri = GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI
+        flow.redirect_uri = redirect_uri
         authorization_url, _ = flow.authorization_url(
             access_type="offline",
             include_granted_scopes="true",
@@ -219,7 +253,7 @@ async def start_google_drive_web_oauth():
         "client_config": client_config,
         "created_at": int(time.time()),
     }
-    REDIS_CONN.set_obj(_web_state_cache_key(flow_id), cache_payload, WEB_FLOW_TTL_SECS)
+    REDIS_CONN.set_obj(_web_state_cache_key(flow_id, source), cache_payload, WEB_FLOW_TTL_SECS)
 
     return get_json_result(
         data={
@@ -230,60 +264,122 @@ async def start_google_drive_web_oauth():
     )
 
 
-@manager.route("/google-drive/oauth/web/callback", methods=["GET"])  # noqa: F821
+@manager.route("/gmail/oauth/web/callback", methods=["GET"])  # noqa: F821
-async def google_drive_web_oauth_callback():
+async def google_gmail_web_oauth_callback():
     state_id = request.args.get("state")
     error = request.args.get("error")
+    source = "gmail"
+    if source != 'gmail':
+        return await _render_web_oauth_popup("", False, "Invalid Google OAuth type.", source)
+
     error_description = request.args.get("error_description") or error
 
     if not state_id:
-        return await _render_web_oauth_popup("", False, "Missing OAuth state parameter.")
+        return await _render_web_oauth_popup("", False, "Missing OAuth state parameter.", source)
 
-    state_cache = REDIS_CONN.get(_web_state_cache_key(state_id))
+    state_cache = REDIS_CONN.get(_web_state_cache_key(state_id, source))
     if not state_cache:
-        return await _render_web_oauth_popup(state_id, False, "Authorization session expired. Please restart from the main window.")
+        return await _render_web_oauth_popup(state_id, False, "Authorization session expired. Please restart from the main window.", source)
 
     state_obj = json.loads(state_cache)
     client_config = state_obj.get("client_config")
     if not client_config:
-        REDIS_CONN.delete(_web_state_cache_key(state_id))
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
-        return await _render_web_oauth_popup(state_id, False, "Authorization session was invalid. Please retry.")
+        return await _render_web_oauth_popup(state_id, False, "Authorization session was invalid. Please retry.", source)
 
     if error:
-        REDIS_CONN.delete(_web_state_cache_key(state_id))
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
-        return await _render_web_oauth_popup(state_id, False, error_description or "Authorization was cancelled.")
+        return await _render_web_oauth_popup(state_id, False, error_description or "Authorization was cancelled.", source)
 
     code = request.args.get("code")
     if not code:
-        return await _render_web_oauth_popup(state_id, False, "Missing authorization code from Google.")
+        return await _render_web_oauth_popup(state_id, False, "Missing authorization code from Google.", source)
 
     try:
-        flow = Flow.from_client_config(client_config, scopes=GOOGLE_SCOPES[DocumentSource.GOOGLE_DRIVE])
+        # TODO(google-oauth): branch scopes/redirect_uri based on source_type (drive vs gmail)
-        flow.redirect_uri = GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI
+        flow = Flow.from_client_config(client_config, scopes=GOOGLE_SCOPES[DocumentSource.GMAIL])
+        flow.redirect_uri = GMAIL_WEB_OAUTH_REDIRECT_URI
         flow.fetch_token(code=code)
     except Exception as exc:  # pragma: no cover - defensive
         logging.exception("Failed to exchange Google OAuth code: %s", exc)
-        REDIS_CONN.delete(_web_state_cache_key(state_id))
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
-        return await _render_web_oauth_popup(state_id, False, "Failed to exchange tokens with Google. Please retry.")
+        return await _render_web_oauth_popup(state_id, False, "Failed to exchange tokens with Google. Please retry.", source)
 
     creds_json = flow.credentials.to_json()
     result_payload = {
         "user_id": state_obj.get("user_id"),
         "credentials": creds_json,
     }
-    REDIS_CONN.set_obj(_web_result_cache_key(state_id), result_payload, WEB_FLOW_TTL_SECS)
+    REDIS_CONN.set_obj(_web_result_cache_key(state_id, source), result_payload, WEB_FLOW_TTL_SECS)
-    REDIS_CONN.delete(_web_state_cache_key(state_id))
 
-    return await _render_web_oauth_popup(state_id, True, "Authorization completed successfully.")
+    print("\n\n", _web_result_cache_key(state_id, source), "\n\n")
+
+    REDIS_CONN.delete(_web_state_cache_key(state_id, source))
+
+    return await _render_web_oauth_popup(state_id, True, "Authorization completed successfully.", source)
 
 
-@manager.route("/google-drive/oauth/web/result", methods=["POST"])  # noqa: F821
+@manager.route("/google-drive/oauth/web/callback", methods=["GET"])  # noqa: F821
+async def google_drive_web_oauth_callback():
+    state_id = request.args.get("state")
+    error = request.args.get("error")
+    source = "google-drive"
+    if source not in ("google-drive", "gmail"):
+        return await _render_web_oauth_popup("", False, "Invalid Google OAuth type.", source)
+
+    error_description = request.args.get("error_description") or error
+
+    if not state_id:
+        return await _render_web_oauth_popup("", False, "Missing OAuth state parameter.", source)
+
+    state_cache = REDIS_CONN.get(_web_state_cache_key(state_id, source))
+    if not state_cache:
+        return await _render_web_oauth_popup(state_id, False, "Authorization session expired. Please restart from the main window.", source)
+
+    state_obj = json.loads(state_cache)
+    client_config = state_obj.get("client_config")
+    if not client_config:
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
+        return await _render_web_oauth_popup(state_id, False, "Authorization session was invalid. Please retry.", source)
+
+    if error:
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
+        return await _render_web_oauth_popup(state_id, False, error_description or "Authorization was cancelled.", source)
+
+    code = request.args.get("code")
+    if not code:
+        return await _render_web_oauth_popup(state_id, False, "Missing authorization code from Google.", source)
+
+    try:
+        # TODO(google-oauth): branch scopes/redirect_uri based on source_type (drive vs gmail)
+        flow = Flow.from_client_config(client_config, scopes=GOOGLE_SCOPES[DocumentSource.GOOGLE_DRIVE])
+        flow.redirect_uri = GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI
+        flow.fetch_token(code=code)
+    except Exception as exc:  # pragma: no cover - defensive
+        logging.exception("Failed to exchange Google OAuth code: %s", exc)
+        REDIS_CONN.delete(_web_state_cache_key(state_id, source))
+        return await _render_web_oauth_popup(state_id, False, "Failed to exchange tokens with Google. Please retry.", source)
+
+    creds_json = flow.credentials.to_json()
+    result_payload = {
+        "user_id": state_obj.get("user_id"),
+        "credentials": creds_json,
+    }
+    REDIS_CONN.set_obj(_web_result_cache_key(state_id, source), result_payload, WEB_FLOW_TTL_SECS)
+    REDIS_CONN.delete(_web_state_cache_key(state_id, source))
+
+    return await _render_web_oauth_popup(state_id, True, "Authorization completed successfully.", source)
+
+@manager.route("/google/oauth/web/result", methods=["POST"])  # noqa: F821
 @login_required
 @validate_request("flow_id")
-async def poll_google_drive_web_result():
+async def poll_google_web_result():
     req = await request.json or {}
+    source = request.args.get("type")
+    if source not in ("google-drive", "gmail"):
+        return get_json_result(code=RetCode.ARGUMENT_ERROR, message="Invalid Google OAuth type.")
     flow_id = req.get("flow_id")
-    cache_raw = REDIS_CONN.get(_web_result_cache_key(flow_id))
+    cache_raw = REDIS_CONN.get(_web_result_cache_key(flow_id, source))
     if not cache_raw:
         return get_json_result(code=RetCode.RUNNING, message="Authorization is still pending.")
 
@@ -291,5 +387,5 @@ async def poll_google_drive_web_result():
     if result.get("user_id") != current_user.id:
         return get_json_result(code=RetCode.PERMISSION_ERROR, message="You are not allowed to access this authorization result.")
 
-    REDIS_CONN.delete(_web_result_cache_key(flow_id))
+    REDIS_CONN.delete(_web_result_cache_key(flow_id, source))
     return get_json_result(data={"credentials": result.get("credentials")})

api/apps/user_app.py

@@ -121,8 +121,8 @@ async def login():
         response_data = user.to_json()
         user.access_token = get_uuid()
         login_user(user)
-        user.update_time = (current_timestamp(),)
+        user.update_time = current_timestamp()
-        user.update_date = (datetime_format(datetime.now()),)
+        user.update_date = datetime_format(datetime.now())
         user.save()
         msg = "Welcome back!"
 
@@ -1002,8 +1002,8 @@ async def forget():
     # Auto login (reuse login flow)
     user.access_token = get_uuid()
     login_user(user)
-    user.update_time = (current_timestamp(),)
+    user.update_time = current_timestamp()
-    user.update_date = (datetime_format(datetime.now()),)
+    user.update_date = datetime_format(datetime.now())
     user.save()
     msg = "Password reset successful. Logged in."
     return construct_response(data=user.to_json(), auth=user.get_id(), message=msg)

common/data_source/config.py

@@ -217,6 +217,7 @@ OAUTH_GOOGLE_DRIVE_CLIENT_SECRET = os.environ.get(
     "OAUTH_GOOGLE_DRIVE_CLIENT_SECRET", ""
 )
 GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI = os.environ.get("GOOGLE_DRIVE_WEB_OAUTH_REDIRECT_URI", "http://localhost:9380/v1/connector/google-drive/oauth/web/callback")
+GMAIL_WEB_OAUTH_REDIRECT_URI = os.environ.get("GMAIL_WEB_OAUTH_REDIRECT_URI", "http://localhost:9380/v1/connector/gmail/oauth/web/callback")
 
 CONFLUENCE_OAUTH_TOKEN_URL = "https://auth.atlassian.com/oauth/token"
 RATE_LIMIT_MESSAGE_LOWERCASE = "Rate limit exceeded".lower()

common/data_source/gmail_connector.py

@@ -1,6 +1,6 @@
 import logging
+import os
 from typing import Any
-
 from google.oauth2.credentials import Credentials as OAuthCredentials
 from google.oauth2.service_account import Credentials as ServiceAccountCredentials
 from googleapiclient.errors import HttpError
@@ -9,10 +9,10 @@ from common.data_source.config import INDEX_BATCH_SIZE, SLIM_BATCH_SIZE, Documen
 from common.data_source.google_util.auth import get_google_creds
 from common.data_source.google_util.constant import DB_CREDENTIALS_PRIMARY_ADMIN_KEY, MISSING_SCOPES_ERROR_STR, SCOPE_INSTRUCTIONS, USER_FIELDS
 from common.data_source.google_util.resource import get_admin_service, get_gmail_service
-from common.data_source.google_util.util import _execute_single_retrieval, execute_paginated_retrieval
+from common.data_source.google_util.util import _execute_single_retrieval, execute_paginated_retrieval, sanitize_filename, clean_string
 from common.data_source.interfaces import LoadConnector, PollConnector, SecondsSinceUnixEpoch, SlimConnectorWithPermSync
 from common.data_source.models import BasicExpertInfo, Document, ExternalAccess, GenerateDocumentsOutput, GenerateSlimDocumentOutput, SlimDocument, TextSection
-from common.data_source.utils import build_time_range_query, clean_email_and_extract_name, get_message_body, is_mail_service_disabled_error, time_str_to_utc
+from common.data_source.utils import build_time_range_query, clean_email_and_extract_name, get_message_body, is_mail_service_disabled_error, gmail_time_str_to_utc
 
 # Constants for Gmail API fields
 THREAD_LIST_FIELDS = "nextPageToken, threads(id)"
@@ -67,7 +67,6 @@ def message_to_section(message: dict[str, Any]) -> tuple[TextSection, dict[str,
             message_data += f"{name}: {value}\n"
 
     message_body_text: str = get_message_body(payload)
-
     return TextSection(link=link, text=message_body_text + message_data), metadata
 
 
@@ -97,13 +96,15 @@ def thread_to_document(full_thread: dict[str, Any], email_used_to_fetch_thread:
 
         if not semantic_identifier:
             semantic_identifier = message_metadata.get("subject", "")
+            semantic_identifier = clean_string(semantic_identifier)
+            semantic_identifier = sanitize_filename(semantic_identifier)
 
         if message_metadata.get("updated_at"):
             updated_at = message_metadata.get("updated_at")
             
     updated_at_datetime = None
     if updated_at:
-        updated_at_datetime = time_str_to_utc(updated_at)
+        updated_at_datetime = gmail_time_str_to_utc(updated_at)
 
     thread_id = full_thread.get("id")
     if not thread_id:
@@ -115,15 +116,24 @@ def thread_to_document(full_thread: dict[str, Any], email_used_to_fetch_thread:
     if not semantic_identifier:
         semantic_identifier = "(no subject)"
 
+    combined_sections = "\n\n".join(
+        sec.text for sec in sections if hasattr(sec, "text")
+    )
+    blob = combined_sections
+    size_bytes = len(blob)
+    extension = '.txt'
+
     return Document(
         id=thread_id,
         semantic_identifier=semantic_identifier,
-        sections=sections,
+        blob=blob,
+        size_bytes=size_bytes,
+        extension=extension,
         source=DocumentSource.GMAIL,
         primary_owners=primary_owners,
         secondary_owners=secondary_owners,
         doc_updated_at=updated_at_datetime,
-        metadata={},
+        metadata=message_metadata,
         external_access=ExternalAccess(
             external_user_emails={email_used_to_fetch_thread},
             external_user_group_ids=set(),
@@ -214,15 +224,13 @@ class GmailConnector(LoadConnector, PollConnector, SlimConnectorWithPermSync):
                     q=query,
                     continue_on_404_or_403=True,
                 ):
-                    full_threads = _execute_single_retrieval(
+                    full_thread = _execute_single_retrieval(
                         retrieval_function=gmail_service.users().threads().get,
-                        list_key=None,
                         userId=user_email,
                         fields=THREAD_FIELDS,
                         id=thread["id"],
                         continue_on_404_or_403=True,
                     )
-                    full_thread = list(full_threads)[0]
                     doc = thread_to_document(full_thread, user_email)
                     if doc is None:
                         continue
@@ -310,4 +318,30 @@ class GmailConnector(LoadConnector, PollConnector, SlimConnectorWithPermSync):
 
 
 if __name__ == "__main__":
-    pass
+    import time
+    import os
+    from common.data_source.google_util.util import get_credentials_from_env
+    logging.basicConfig(level=logging.INFO)
+    try:
+        email = os.environ.get("GMAIL_TEST_EMAIL", "newyorkupperbay@gmail.com")
+        creds = get_credentials_from_env(email, oauth=True, source="gmail")
+        print("Credentials loaded successfully")
+        print(f"{creds=}")
+
+        connector = GmailConnector(batch_size=2)
+        print("GmailConnector initialized")
+        connector.load_credentials(creds)
+        print("Credentials loaded into connector")
+
+        print("Gmail is ready to use")
+
+        for file in connector._fetch_threads(
+            int(time.time()) - 1 * 24 * 60 * 60,
+            int(time.time()),
+        ):
+            print("new batch","-"*80)
+            for f in file:
+                print(f)
+                print("\n\n")
+    except Exception as e:
+        logging.exception(f"Error loading credentials: {e}")

common/data_source/google_drive/connector.py

@@ -1,7 +1,6 @@
 """Google Drive connector"""
 
 import copy
-import json
 import logging
 import os
 import sys
@@ -32,7 +31,6 @@ from common.data_source.google_drive.file_retrieval import (
 from common.data_source.google_drive.model import DriveRetrievalStage, GoogleDriveCheckpoint, GoogleDriveFileType, RetrievedDriveFile, StageCompletion
 from common.data_source.google_util.auth import get_google_creds
 from common.data_source.google_util.constant import DB_CREDENTIALS_PRIMARY_ADMIN_KEY, MISSING_SCOPES_ERROR_STR, USER_FIELDS
-from common.data_source.google_util.oauth_flow import ensure_oauth_token_dict
 from common.data_source.google_util.resource import GoogleDriveService, get_admin_service, get_drive_service
 from common.data_source.google_util.util import GoogleFields, execute_paginated_retrieval, get_file_owners
 from common.data_source.google_util.util_threadpool_concurrency import ThreadSafeDict
@@ -1138,39 +1136,6 @@ class GoogleDriveConnector(SlimConnectorWithPermSync, CheckpointedConnectorWithP
         return GoogleDriveCheckpoint.model_validate_json(checkpoint_json)
 
 
-def get_credentials_from_env(email: str, oauth: bool = False) -> dict:
-    try:
-        if oauth:
-            raw_credential_string = os.environ["GOOGLE_DRIVE_OAUTH_CREDENTIALS_JSON_STR"]
-        else:
-            raw_credential_string = os.environ["GOOGLE_DRIVE_SERVICE_ACCOUNT_JSON_STR"]
-    except KeyError:
-        raise ValueError("Missing Google Drive credentials in environment variables")
-
-    try:
-        credential_dict = json.loads(raw_credential_string)
-    except json.JSONDecodeError:
-        raise ValueError("Invalid JSON in Google Drive credentials")
-
-    if oauth:
-        credential_dict = ensure_oauth_token_dict(credential_dict, DocumentSource.GOOGLE_DRIVE)
-
-    refried_credential_string = json.dumps(credential_dict)
-
-    DB_CREDENTIALS_DICT_TOKEN_KEY = "google_tokens"
-    DB_CREDENTIALS_DICT_SERVICE_ACCOUNT_KEY = "google_service_account_key"
-    DB_CREDENTIALS_PRIMARY_ADMIN_KEY = "google_primary_admin"
-    DB_CREDENTIALS_AUTHENTICATION_METHOD = "authentication_method"
-
-    cred_key = DB_CREDENTIALS_DICT_TOKEN_KEY if oauth else DB_CREDENTIALS_DICT_SERVICE_ACCOUNT_KEY
-
-    return {
-        cred_key: refried_credential_string,
-        DB_CREDENTIALS_PRIMARY_ADMIN_KEY: email,
-        DB_CREDENTIALS_AUTHENTICATION_METHOD: "uploaded",
-    }
-
-
 class CheckpointOutputWrapper:
     """
     Wraps a CheckpointOutput generator to give things back in a more digestible format.
@@ -1236,7 +1201,7 @@ def yield_all_docs_from_checkpoint_connector(
 
 if __name__ == "__main__":
     import time
-
+    from common.data_source.google_util.util import get_credentials_from_env
     logging.basicConfig(level=logging.DEBUG)
 
     try:
@@ -1245,7 +1210,7 @@ if __name__ == "__main__":
         creds = get_credentials_from_env(email, oauth=True)
         print("Credentials loaded successfully")
         print(f"{creds=}")
-
+        sys.exit(0)
         connector = GoogleDriveConnector(
             include_shared_drives=False,
             shared_drive_urls=None,

common/data_source/google_util/constant.py

@@ -49,11 +49,11 @@ MISSING_SCOPES_ERROR_STR = "client not authorized for any of the scopes requeste
 SCOPE_INSTRUCTIONS = ""
 
 
-GOOGLE_DRIVE_WEB_OAUTH_POPUP_TEMPLATE = """<!DOCTYPE html>
+GOOGLE_WEB_OAUTH_POPUP_TEMPLATE = """<!DOCTYPE html>
 <html lang="en">
 <head>
   <meta charset="utf-8" />
-  <title>Google Drive Authorization</title>
+  <title>{title}</title>
   <style>
     body {{
       font-family: Arial, sans-serif;

common/data_source/google_util/util.py

@@ -1,12 +1,17 @@
+import json
 import logging
+import os
+import re
 import socket
 from collections.abc import Callable, Iterator
 from enum import Enum
 from typing import Any
-
+import unicodedata
 from googleapiclient.errors import HttpError  # type: ignore  # type: ignore
 
+from common.data_source.config import DocumentSource
 from common.data_source.google_drive.model import GoogleDriveFileType
+from common.data_source.google_util.oauth_flow import ensure_oauth_token_dict
 
 
 # See https://developers.google.com/drive/api/reference/rest/v3/files/list for more
@@ -117,6 +122,7 @@ def _execute_single_retrieval(
     """Execute a single retrieval from Google Drive API"""
     try:
         results = retrieval_function(**request_kwargs).execute()
+
     except HttpError as e:
         if e.resp.status >= 500:
             results = retrieval_function()
@@ -148,5 +154,110 @@ def _execute_single_retrieval(
             error,
         )
         results = retrieval_function()
-
     return results
+
+
+def get_credentials_from_env(email: str, oauth: bool = False, source="drive") -> dict:
+    try:
+        if oauth:
+            raw_credential_string = os.environ["GOOGLE_OAUTH_CREDENTIALS_JSON_STR"]
+        else:
+            raw_credential_string = os.environ["GOOGLE_SERVICE_ACCOUNT_JSON_STR"]
+    except KeyError:
+        raise ValueError("Missing Google Drive credentials in environment variables")
+
+    try:
+        credential_dict = json.loads(raw_credential_string)
+    except json.JSONDecodeError:
+        raise ValueError("Invalid JSON in Google Drive credentials")
+
+    if oauth and source == "drive":
+        credential_dict = ensure_oauth_token_dict(credential_dict, DocumentSource.GOOGLE_DRIVE)
+    else:
+        credential_dict = ensure_oauth_token_dict(credential_dict, DocumentSource.GMAIL)
+
+    refried_credential_string = json.dumps(credential_dict)
+
+    DB_CREDENTIALS_DICT_TOKEN_KEY = "google_tokens"
+    DB_CREDENTIALS_DICT_SERVICE_ACCOUNT_KEY = "google_service_account_key"
+    DB_CREDENTIALS_PRIMARY_ADMIN_KEY = "google_primary_admin"
+    DB_CREDENTIALS_AUTHENTICATION_METHOD = "authentication_method"
+
+    cred_key = DB_CREDENTIALS_DICT_TOKEN_KEY if oauth else DB_CREDENTIALS_DICT_SERVICE_ACCOUNT_KEY
+
+    return {
+        cred_key: refried_credential_string,
+        DB_CREDENTIALS_PRIMARY_ADMIN_KEY: email,
+        DB_CREDENTIALS_AUTHENTICATION_METHOD: "uploaded",
+    }
+
+def sanitize_filename(name: str) -> str:
+    """
+    Soft sanitize for MinIO/S3:
+    - Replace only prohibited characters with a space.
+    - Preserve readability (no ugly underscores).
+    - Collapse multiple spaces.
+    """
+    if name is None:
+        return "file.txt"
+
+    name = str(name).strip()
+
+    # Characters that MUST NOT appear in S3/MinIO object keys
+    # Replace them with a space (not underscore)
+    forbidden = r'[\\\?\#\%\*\:\|\<\>"]'
+    name = re.sub(forbidden, " ", name)
+
+    # Replace slashes "/" (S3 interprets as folder) with space
+    name = name.replace("/", " ")
+
+    # Collapse multiple spaces into one
+    name = re.sub(r"\s+", " ", name)
+
+    # Trim both ends
+    name = name.strip()
+
+    # Enforce reasonable max length
+    if len(name) > 200:
+        base, ext = os.path.splitext(name)
+        name = base[:180].rstrip() + ext
+
+    # Ensure there is an extension (your original logic)
+    if not os.path.splitext(name)[1]:
+        name += ".txt"
+
+    return name
+
+
+def clean_string(text: str | None) -> str | None:
+    """
+    Clean a string to make it safe for insertion into MySQL (utf8mb4).
+    - Normalize Unicode
+    - Remove control characters / zero-width characters
+    - Optionally remove high-plane emoji and symbols
+    """
+    if text is None:
+        return None
+
+    # 0. Ensure the value is a string
+    text = str(text)
+
+    # 1. Normalize Unicode (NFC)
+    text = unicodedata.normalize("NFC", text)
+
+    # 2. Remove ASCII control characters (except tab, newline, carriage return)
+    text = re.sub(r"[\x00-\x08\x0b\x0c\x0e-\x1f\x7f]", "", text)
+
+    # 3. Remove zero-width characters / BOM
+    text = re.sub(r"[\u200b-\u200d\uFEFF]", "", text)
+
+    # 4. Remove high Unicode characters (emoji, special symbols)
+    text = re.sub(r"[\U00010000-\U0010FFFF]", "", text)
+
+    # 5. Final fallback: strip any invalid UTF-8 sequences
+    try:
+        text.encode("utf-8")
+    except UnicodeEncodeError:
+        text = text.encode("utf-8", errors="ignore").decode("utf-8")
+
+    return text

common/data_source/interfaces.py

@@ -30,7 +30,6 @@ class LoadConnector(ABC):
         """Load documents from state"""
         pass
 
-    @abstractmethod
     def validate_connector_settings(self) -> None:
         """Validate connector settings"""
         pass

common/data_source/utils.py

@@ -733,7 +733,7 @@ def build_time_range_query(
     """Build time range query for Gmail API"""
     query = ""
     if time_range_start is not None and time_range_start != 0:
-        query += f"after:{int(time_range_start)}"
+        query += f"after:{int(time_range_start) + 1}"
     if time_range_end is not None and time_range_end != 0:
         query += f" before:{int(time_range_end)}"
     query = query.strip()
@@ -778,6 +778,15 @@ def time_str_to_utc(time_str: str):
     return datetime.fromisoformat(time_str.replace("Z", "+00:00"))
 
 
+def gmail_time_str_to_utc(time_str: str):
+    """Convert Gmail RFC 2822 time string to UTC."""
+    from email.utils import parsedate_to_datetime
+    from datetime import timezone
+
+    dt = parsedate_to_datetime(time_str)
+    return dt.astimezone(timezone.utc)
+
+
 # Notion Utilities
 T = TypeVar("T")
 

rag/svr/sync_data_source.py

@@ -41,6 +41,7 @@ from common.data_source import BlobStorageConnector, NotionConnector, DiscordCon
 from common.constants import FileSource, TaskStatus
 from common.data_source.config import INDEX_BATCH_SIZE
 from common.data_source.confluence_connector import ConfluenceConnector
+from common.data_source.gmail_connector import GmailConnector
 from common.data_source.interfaces import CheckpointOutputWrapper
 from common.data_source.utils import load_all_docs_from_checkpoint_connector
 from common.log_utils import init_root_logger
@@ -230,7 +231,64 @@ class Gmail(SyncBase):
     SOURCE_NAME: str = FileSource.GMAIL
 
     async def _generate(self, task: dict):
-        pass
+        # Gmail sync reuses the generic LoadConnector/PollConnector interface
+        # implemented by common.data_source.gmail_connector.GmailConnector.
+        #
+        # Config expectations (self.conf):
+        #   credentials: Gmail / Workspace OAuth JSON (with primary admin email)
+        #   batch_size:  optional, defaults to INDEX_BATCH_SIZE
+        batch_size = self.conf.get("batch_size", INDEX_BATCH_SIZE)
+
+        self.connector = GmailConnector(batch_size=batch_size)
+
+        credentials = self.conf.get("credentials")
+        if not credentials:
+            raise ValueError("Gmail connector is missing credentials.")
+
+        new_credentials = self.connector.load_credentials(credentials)
+        if new_credentials:
+            # Persist rotated / refreshed credentials back to connector config
+            try:
+                updated_conf = copy.deepcopy(self.conf)
+                updated_conf["credentials"] = new_credentials
+                ConnectorService.update_by_id(task["connector_id"], {"config": updated_conf})
+                self.conf = updated_conf
+                logging.info(
+                    "Persisted refreshed Gmail credentials for connector %s",
+                    task["connector_id"],
+                )
+            except Exception:
+                logging.exception(
+                    "Failed to persist refreshed Gmail credentials for connector %s",
+                    task["connector_id"],
+                )
+
+        # Decide between full reindex and incremental polling by time range.
+        if task["reindex"] == "1" or not task.get("poll_range_start"):
+            start_time = None
+            end_time = None
+            begin_info = "totally"
+            document_generator = self.connector.load_from_state()
+        else:
+            poll_start = task["poll_range_start"]
+            # Defensive: if poll_start is somehow None, fall back to full load
+            if poll_start is None:
+                start_time = None
+                end_time = None
+                begin_info = "totally"
+                document_generator = self.connector.load_from_state()
+            else:
+                start_time = poll_start.timestamp()
+                end_time = datetime.now(timezone.utc).timestamp()
+                begin_info = f"from {poll_start}"
+                document_generator = self.connector.poll_source(start_time, end_time)
+
+        try:
+            admin_email = self.connector.primary_admin_email
+        except RuntimeError:
+            admin_email = "unknown"
+        logging.info(f"Connect to Gmail as {admin_email} {begin_info}")
+        return document_generator
 
 
 class Dropbox(SyncBase):

web/src/assets/svg/data-source/gmail.svg

@@ -0,0 +1,7 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="52 42 88 66">
+<path fill="#4285f4" d="M58 108h14V74L52 59v43c0 3.32 2.69 6 6 6"/>
+<path fill="#34a853" d="M120 108h14c3.32 0 6-2.69 6-6V59l-20 15"/>
+<path fill="#fbbc04" d="M120 48v26l20-15v-8c0-7.42-8.47-11.65-14.4-7.2"/>
+<path fill="#ea4335" d="M72 74V48l24 18 24-18v26L96 92"/>
+<path fill="#c5221f" d="M52 51v8l20 15V48l-5.6-4.2c-5.94-4.45-14.4-.22-14.4 7.2"/>
+</svg>

web/src/locales/en.ts

@@ -739,6 +739,7 @@ Example: Virtual Hosted Style`,
         'Sync pages and databases from Notion for knowledge retrieval.',
       google_driveDescription:
         'Connect your Google Drive via OAuth and sync specific folders or drives.',
+      gmailDescription: 'Connect your Gmail via OAuth to sync emails.',
       webdavDescription: 'Connect to WebDAV servers to sync files.',
       webdavRemotePathTip:
         'Optional: Specify a folder path on the WebDAV server (e.g., /Documents). Leave empty to sync from root.',
@@ -750,6 +751,10 @@ Example: Virtual Hosted Style`,
         'Comma-separated emails whose "My Drive" contents should be indexed (include the primary admin).',
       google_driveSharedFoldersTip:
         'Comma-separated Google Drive folder links to crawl.',
+      gmailPrimaryAdminTip:
+        'Primary admin email with Gmail / Workspace access, used to enumerate domain users and as the default sync account.',
+      gmailTokenTip:
+        'Upload the OAuth JSON generated from Google Console. If it only contains client credentials, run the browser-based verification once to mint long-lived refresh tokens.',
       dropboxDescription:
         'Connect your Dropbox to sync files and folders from a chosen account.',
       dropboxAccessTokenTip:

web/src/locales/ru.ts

@@ -736,6 +736,8 @@ export default {
         'Синхронизируйте страницы и базы данных из Notion для извлечения знаний.',
       google_driveDescription:
         'Подключите ваш Google Drive через OAuth и синхронизируйте определенные папки или диски.',
+      gmailDescription:
+        'Подключите ваш Gmail / Google Workspace аккаунт для синхронизации писем и их метаданных, чтобы построить корпоративную почтовую базу знаний и поиск с учетом прав доступа.',
       google_driveTokenTip:
         'Загрузите JSON токена OAuth, сгенерированный из помощника OAuth или Google Cloud Console. Вы также можете загрузить client_secret JSON из "установленного" или "веб" приложения. Если это ваша первая синхронизация, откроется окно браузера для завершения согласия OAuth. Если JSON уже содержит токен обновления, он будет автоматически повторно использован.',
       google_drivePrimaryAdminTip:
@@ -744,6 +746,10 @@ export default {
         'Электронные почты через запятую, чье содержимое "Мой диск" должно индексироваться (включите основного администратора).',
       google_driveSharedFoldersTip:
         'Ссылки на папки Google Drive через запятую для обхода.',
+      gmailPrimaryAdminTip:
+        'Основной административный email с доступом к Gmail / Workspace, используется для перечисления пользователей домена и как аккаунт синхронизации по умолчанию.',
+      gmailTokenTip:
+        'Загрузите OAuth JSON, сгенерированный в Google Console. Если он содержит только учетные данные клиента, выполните одноразовое подтверждение в браузере, чтобы получить долгоживущие токены обновления.',
       jiraDescription:
         'Подключите ваше рабочее пространство Jira для синхронизации задач, комментариев и вложений.',
       jiraBaseUrlTip:

web/src/locales/zh.ts

@@ -718,6 +718,7 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       notionDescription: ' 同步 Notion 页面与数据库，用于知识检索。',
       google_driveDescription:
         '通过 OAuth 连接 Google Drive，并同步指定的文件夹或云端硬盘。',
+      gmailDescription: '通过 OAuth 连接 Gmail，用于同步邮件。',
       google_driveTokenTip:
         '请上传由 OAuth helper 或 Google Cloud Console 导出的 OAuth token JSON。也支持上传 “installed” 或 “web” 类型的 client_secret JSON。若为首次同步，将自动弹出浏览器完成 OAuth 授权流程；如果该 JSON 已包含 refresh token，将会被自动复用。',
       google_drivePrimaryAdminTip: '拥有相应 Drive 访问权限的管理员邮箱。',
@@ -725,6 +726,10 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
         '需要索引其 “我的云端硬盘” 的邮箱，多个邮箱用逗号分隔（建议包含管理员）。',
       google_driveSharedFoldersTip:
         '需要同步的 Google Drive 文件夹链接，多个链接用逗号分隔。',
+      gmailPrimaryAdminTip:
+        '拥有 Gmail / Workspace 访问权限的主要管理员邮箱，用于列出域内用户并作为默认同步账号。',
+      gmailTokenTip:
+        '请上传由 Google Console 生成的 OAuth JSON。如果仅包含 client credentials，请通过浏览器授权一次以获取长期有效的刷新 Token。',
       dropboxDescription: '连接 Dropbox，同步指定账号下的文件与文件夹。',
       dropboxAccessTokenTip:
         '请在 Dropbox App Console 生成 Access Token，并勾选 files.metadata.read、files.content.read、sharing.read 等必要权限。',

web/src/pages/user-setting/data-source/add-datasource-modal.tsx

@@ -47,6 +47,7 @@ const AddDataSourceModal = ({
       }
       open={visible || false}
       onOpenChange={(open) => !open && hideModal?.()}
+      maskClosable={false}
       // onOk={() => handleOk()}
       okText={t('common.confirm')}
       cancelText={t('common.cancel')}

web/src/pages/user-setting/data-source/component/gmail-token-field.tsx

@@ -0,0 +1,391 @@
+import { useCallback, useEffect, useMemo, useRef, useState } from 'react';
+
+import { FileUploader } from '@/components/file-uploader';
+import { Button } from '@/components/ui/button';
+import {
+  Dialog,
+  DialogContent,
+  DialogDescription,
+  DialogFooter,
+  DialogHeader,
+  DialogTitle,
+} from '@/components/ui/dialog';
+import message from '@/components/ui/message';
+import { FileMimeType } from '@/constants/common';
+import {
+  pollGmailWebAuthResult,
+  startGmailWebAuth,
+} from '@/services/data-source-service';
+import { Loader2 } from 'lucide-react';
+
+export type GmailTokenFieldProps = {
+  value?: string;
+  onChange: (value: any) => void;
+  placeholder?: string;
+};
+
+const credentialHasRefreshToken = (content: string) => {
+  try {
+    const parsed = JSON.parse(content);
+    return Boolean(parsed?.refresh_token);
+  } catch {
+    return false;
+  }
+};
+
+const describeCredentials = (content?: string) => {
+  if (!content) return '';
+  try {
+    const parsed = JSON.parse(content);
+    if (parsed?.refresh_token) {
+      return 'Uploaded OAuth tokens with a refresh token.';
+    }
+    if (parsed?.installed || parsed?.web) {
+      return 'Client credentials detected. Complete verification to mint long-lived tokens.';
+    }
+    return 'Stored Google credential JSON.';
+  } catch {
+    return '';
+  }
+};
+
+const GmailTokenField = ({
+  value,
+  onChange,
+  placeholder,
+}: GmailTokenFieldProps) => {
+  const [files, setFiles] = useState<File[]>([]);
+  const [pendingCredentials, setPendingCredentials] = useState<string>('');
+  const [dialogOpen, setDialogOpen] = useState(false);
+  const [webAuthLoading, setWebAuthLoading] = useState(false);
+  const [webFlowId, setWebFlowId] = useState<string | null>(null);
+  const [webStatus, setWebStatus] = useState<
+    'idle' | 'waiting' | 'success' | 'error'
+  >('idle');
+  const [webStatusMessage, setWebStatusMessage] = useState('');
+  const webFlowIdRef = useRef<string | null>(null);
+  const webPollTimerRef = useRef<ReturnType<typeof setTimeout> | null>(null);
+
+  const clearWebState = useCallback(() => {
+    if (webPollTimerRef.current) {
+      clearTimeout(webPollTimerRef.current);
+      webPollTimerRef.current = null;
+    }
+    webFlowIdRef.current = null;
+    setWebFlowId(null);
+    setWebStatus('idle');
+    setWebStatusMessage('');
+  }, []);
+
+  useEffect(() => {
+    return () => {
+      if (webPollTimerRef.current) {
+        clearTimeout(webPollTimerRef.current);
+      }
+    };
+  }, []);
+
+  useEffect(() => {
+    webFlowIdRef.current = webFlowId;
+  }, [webFlowId]);
+
+  const credentialSummary = useMemo(() => describeCredentials(value), [value]);
+  const hasVerifiedTokens = useMemo(
+    () => Boolean(value && credentialHasRefreshToken(value)),
+    [value],
+  );
+  const hasUploadedButUnverified = useMemo(
+    () => Boolean(value && !hasVerifiedTokens),
+    [hasVerifiedTokens, value],
+  );
+
+  const resetDialog = useCallback(
+    (shouldResetState: boolean) => {
+      setDialogOpen(false);
+      clearWebState();
+      if (shouldResetState) {
+        setPendingCredentials('');
+        setFiles([]);
+      }
+    },
+    [clearWebState],
+  );
+
+  const fetchWebResult = useCallback(
+    async (flowId: string) => {
+      try {
+        const { data } = await pollGmailWebAuthResult({
+          flow_id: flowId,
+        });
+        if (data.code === 0 && data.data?.credentials) {
+          onChange(data.data.credentials);
+          setPendingCredentials('');
+          message.success('Gmail credentials verified.');
+          resetDialog(false);
+          return;
+        }
+        if (data.code === 106) {
+          setWebStatus('waiting');
+          setWebStatusMessage('Authorization confirmed. Finalizing tokens...');
+          if (webPollTimerRef.current) {
+            clearTimeout(webPollTimerRef.current);
+          }
+          webPollTimerRef.current = setTimeout(
+            () => fetchWebResult(flowId),
+            1500,
+          );
+          return;
+        }
+        message.error(data.message || 'Authorization failed.');
+        clearWebState();
+      } catch (err) {
+        message.error('Unable to retrieve authorization result.');
+        clearWebState();
+      }
+    },
+    [clearWebState, onChange, resetDialog],
+  );
+
+  useEffect(() => {
+    const handler = (event: MessageEvent) => {
+      const payload = event.data;
+      if (!payload || payload.type !== 'ragflow-gmail-oauth') {
+        return;
+      }
+      if (!payload.flowId) {
+        return;
+      }
+      if (webFlowIdRef.current && webFlowIdRef.current !== payload.flowId) {
+        return;
+      }
+
+      if (payload.status === 'success') {
+        setWebStatus('waiting');
+        setWebStatusMessage('Authorization confirmed. Finalizing tokens...');
+        fetchWebResult(payload.flowId);
+      } else {
+        message.error(
+          payload.message || 'Authorization window reported an error.',
+        );
+        clearWebState();
+      }
+    };
+
+    window.addEventListener('message', handler);
+    return () => window.removeEventListener('message', handler);
+  }, [clearWebState, fetchWebResult]);
+
+  const handleValueChange = useCallback(
+    (nextFiles: File[]) => {
+      if (!nextFiles.length) {
+        setFiles([]);
+        onChange('');
+        setPendingCredentials('');
+        clearWebState();
+        return;
+      }
+      const file = nextFiles[nextFiles.length - 1];
+      file
+        .text()
+        .then((text) => {
+          try {
+            JSON.parse(text);
+          } catch {
+            message.error('Invalid JSON file.');
+            setFiles([]);
+            clearWebState();
+            return;
+          }
+          setFiles([file]);
+          clearWebState();
+          if (credentialHasRefreshToken(text)) {
+            onChange(text);
+            setPendingCredentials('');
+            message.success('Gmail OAuth credentials uploaded.');
+            return;
+          }
+          setPendingCredentials(text);
+          setDialogOpen(true);
+          message.info(
+            'Client configuration uploaded. Verification is required to finish setup.',
+          );
+        })
+        .catch(() => {
+          message.error('Unable to read the uploaded file.');
+          setFiles([]);
+        });
+    },
+    [clearWebState, onChange],
+  );
+
+  const handleStartWebAuthorization = useCallback(async () => {
+    if (!pendingCredentials) {
+      message.error('No Google credential file detected.');
+      return;
+    }
+    setWebAuthLoading(true);
+    clearWebState();
+    try {
+      const { data } = await startGmailWebAuth({
+        credentials: pendingCredentials,
+      });
+      if (data.code === 0 && data.data?.authorization_url) {
+        const flowId = data.data.flow_id;
+        const popup = window.open(
+          data.data.authorization_url,
+          'ragflow-gmail-oauth',
+          'width=600,height=720',
+        );
+        if (!popup) {
+          message.error(
+            'Popup was blocked. Please allow popups for this site.',
+          );
+          return;
+        }
+        popup.focus();
+        webFlowIdRef.current = flowId;
+        setWebFlowId(flowId);
+        setWebStatus('waiting');
+        setWebStatusMessage('Complete the Google consent in the popup window.');
+      } else {
+        message.error(data.message || 'Failed to start browser authorization.');
+      }
+    } catch (err) {
+      message.error('Failed to start browser authorization.');
+    } finally {
+      setWebAuthLoading(false);
+    }
+  }, [clearWebState, pendingCredentials]);
+
+  const handleManualWebCheck = useCallback(() => {
+    if (!webFlowId) {
+      message.info('Start browser authorization first.');
+      return;
+    }
+    setWebStatus('waiting');
+    setWebStatusMessage('Checking authorization status...');
+    fetchWebResult(webFlowId);
+  }, [fetchWebResult, webFlowId]);
+
+  const handleCancel = useCallback(() => {
+    message.warning(
+      'Verification canceled. Upload the credential again to restart.',
+    );
+    resetDialog(true);
+  }, [resetDialog]);
+
+  return (
+    <div className="flex flex-col gap-3">
+      {(credentialSummary ||
+        hasVerifiedTokens ||
+        hasUploadedButUnverified ||
+        pendingCredentials) && (
+        <div className="flex flex-wrap items-center gap-3 rounded-md border border-dashed border-muted-foreground/40 bg-muted/20 px-3 py-2 text-xs text-muted-foreground">
+          <div className="flex flex-wrap items-center gap-2">
+            {hasVerifiedTokens ? (
+              <span className="rounded-full bg-emerald-100 px-2 py-0.5 text-[11px] font-semibold uppercase tracking-wide text-emerald-700">
+                Verified
+              </span>
+            ) : null}
+            {hasUploadedButUnverified ? (
+              <span className="rounded-full bg-amber-100 px-2 py-0.5 text-[11px] font-semibold uppercase tracking-wide text-amber-700">
+                Needs authorization
+              </span>
+            ) : null}
+            {pendingCredentials && !hasVerifiedTokens ? (
+              <span className="rounded-full bg-blue-100 px-2 py-0.5 text-[11px] font-semibold uppercase tracking-wide text-blue-700">
+                Uploaded (pending)
+              </span>
+            ) : null}
+          </div>
+          {credentialSummary ? (
+            <p className="m-0">{credentialSummary}</p>
+          ) : null}
+        </div>
+      )}
+      <FileUploader
+        className="py-4 border-[0.5px] bg-bg-card text-text-secondary"
+        value={files}
+        onValueChange={handleValueChange}
+        accept={{ '*.json': [FileMimeType.Json] }}
+        maxFileCount={1}
+        description={'Upload your Gmail OAuth JSON file.'}
+      />
+
+      <Dialog
+        open={dialogOpen}
+        onOpenChange={(open) => {
+          if (!open && dialogOpen) {
+            handleCancel();
+          }
+        }}
+      >
+        <DialogContent
+          onPointerDownOutside={(e) => e.preventDefault()}
+          onInteractOutside={(e) => e.preventDefault()}
+          onEscapeKeyDown={(e) => e.preventDefault()}
+        >
+          <DialogHeader>
+            <DialogTitle>Complete Gmail verification</DialogTitle>
+            <DialogDescription>
+              The uploaded client credentials do not contain a refresh token.
+              Run the verification flow once to mint reusable tokens.
+            </DialogDescription>
+          </DialogHeader>
+
+          <div className="space-y-4">
+            <div className="rounded-md border border-dashed border-muted-foreground/40 bg-muted/10 px-4 py-4 text-sm text-muted-foreground">
+              <div className="text-sm font-semibold text-foreground">
+                Authorize in browser
+              </div>
+              <p className="mt-2">
+                We will open Google&apos;s consent page in a new window. Sign in
+                with the admin account, grant access, and return here. Your
+                credentials will update automatically.
+              </p>
+              {webStatus !== 'idle' && (
+                <p
+                  className={`mt-2 text-xs ${
+                    webStatus === 'error'
+                      ? 'text-destructive'
+                      : 'text-muted-foreground'
+                  }`}
+                >
+                  {webStatusMessage}
+                </p>
+              )}
+              <div className="mt-3 flex flex-wrap gap-2">
+                <Button
+                  onClick={handleStartWebAuthorization}
+                  disabled={webAuthLoading}
+                >
+                  {webAuthLoading && (
+                    <Loader2 className="mr-2 size-4 animate-spin" />
+                  )}
+                  Authorize with Google
+                </Button>
+                {webFlowId ? (
+                  <Button
+                    variant="outline"
+                    onClick={handleManualWebCheck}
+                    disabled={webStatus === 'success'}
+                  >
+                    Refresh status
+                  </Button>
+                ) : null}
+              </div>
+            </div>
+          </div>
+
+          <DialogFooter className="pt-2">
+            <Button variant="ghost" onClick={handleCancel}>
+              Cancel
+            </Button>
+          </DialogFooter>
+        </DialogContent>
+      </Dialog>
+    </div>
+  );
+};
+
+export default GmailTokenField;

web/src/pages/user-setting/data-source/component/google-drive-token-field.tsx

@@ -1,5 +1,3 @@
-import { useCallback, useEffect, useMemo, useRef, useState } from 'react';
-
 import { FileUploader } from '@/components/file-uploader';
 import { Button } from '@/components/ui/button';
 import {
@@ -17,6 +15,7 @@ import {
   startGoogleDriveWebAuth,
 } from '@/services/data-source-service';
 import { Loader2 } from 'lucide-react';
+import { useCallback, useEffect, useMemo, useRef, useState } from 'react';
 
 type GoogleDriveTokenFieldProps = {
   value?: string;
@@ -313,12 +312,16 @@ const GoogleDriveTokenField = ({
       <Dialog
         open={dialogOpen}
         onOpenChange={(open) => {
-          if (!open) {
+          if (!open && dialogOpen) {
             handleCancel();
           }
         }}
       >
-        <DialogContent>
+        <DialogContent
+          onPointerDownOutside={(e) => e.preventDefault()}
+          onInteractOutside={(e) => e.preventDefault()}
+          onEscapeKeyDown={(e) => e.preventDefault()}
+        >
           <DialogHeader>
             <DialogTitle>Complete Google verification</DialogTitle>
             <DialogDescription>
@@ -326,7 +329,6 @@ const GoogleDriveTokenField = ({
               Run the verification flow once to mint reusable tokens.
             </DialogDescription>
           </DialogHeader>
-
           <div className="space-y-4">
             <div className="rounded-md border border-dashed border-muted-foreground/40 bg-muted/10 px-4 py-4 text-sm text-muted-foreground">
               <div className="text-sm font-semibold text-foreground">
@@ -370,7 +372,6 @@ const GoogleDriveTokenField = ({
               </div>
             </div>
           </div>
-
           <DialogFooter className="pt-2">
             <Button variant="ghost" onClick={handleCancel}>
               Cancel

web/src/pages/user-setting/data-source/contant.tsx

@@ -1,6 +1,7 @@
 import { FormFieldType } from '@/components/dynamic-form';
 import SvgIcon from '@/components/svg-icon';
 import { t } from 'i18next';
+import GmailTokenField from './component/gmail-token-field';
 import GoogleDriveTokenField from './component/google-drive-token-field';
 
 export enum DataSourceKey {
@@ -10,7 +11,7 @@ export enum DataSourceKey {
   DISCORD = 'discord',
   GOOGLE_DRIVE = 'google_drive',
   MOODLE = 'moodle',
-  //   GMAIL = 'gmail',
+  GMAIL = 'gmail',
   JIRA = 'jira',
   WEBDAV = 'webdav',
   DROPBOX = 'dropbox',
@@ -45,6 +46,11 @@ export const DataSourceInfo = {
     description: t(`setting.${DataSourceKey.GOOGLE_DRIVE}Description`),
     icon: <SvgIcon name={'data-source/google-drive'} width={38} />,
   },
+  [DataSourceKey.GMAIL]: {
+    name: 'Gmail',
+    description: t(`setting.${DataSourceKey.GMAIL}Description`),
+    icon: <SvgIcon name={'data-source/gmail'} width={38} />,
+  },
   [DataSourceKey.MOODLE]: {
     name: 'Moodle',
     description: t(`setting.${DataSourceKey.MOODLE}Description`),
@@ -320,6 +326,38 @@ export const DataSourceFormFields = {
       defaultValue: 'uploaded',
     },
   ],
+  [DataSourceKey.GMAIL]: [
+    {
+      label: 'Primary Admin Email',
+      name: 'config.credentials.google_primary_admin',
+      type: FormFieldType.Text,
+      required: true,
+      placeholder: 'admin@example.com',
+      tooltip: t('setting.gmailPrimaryAdminTip'),
+    },
+    {
+      label: 'OAuth Token JSON',
+      name: 'config.credentials.google_tokens',
+      type: FormFieldType.Textarea,
+      required: true,
+      render: (fieldProps: any) => (
+        <GmailTokenField
+          value={fieldProps.value}
+          onChange={fieldProps.onChange}
+          placeholder='{ "token": "...", "refresh_token": "...", ... }'
+        />
+      ),
+      tooltip: t('setting.gmailTokenTip'),
+    },
+    {
+      label: '',
+      name: 'config.credentials.authentication_method',
+      type: FormFieldType.Text,
+      required: false,
+      hidden: true,
+      defaultValue: 'uploaded',
+    },
+  ],
   [DataSourceKey.MOODLE]: [
     {
       label: 'Moodle URL',
@@ -550,6 +588,17 @@ export const DataSourceFormDefaultValues = {
       },
     },
   },
+  [DataSourceKey.GMAIL]: {
+    name: '',
+    source: DataSourceKey.GMAIL,
+    config: {
+      credentials: {
+        google_primary_admin: '',
+        google_tokens: '',
+        authentication_method: 'uploaded',
+      },
+    },
+  },
   [DataSourceKey.MOODLE]: {
     name: '',
     source: DataSourceKey.MOODLE,

web/src/services/data-source-service.ts

@@ -34,9 +34,17 @@ export const featchDataSourceDetail = (id: string) =>
   request.get(api.dataSourceDetail(id));
 
 export const startGoogleDriveWebAuth = (payload: { credentials: string }) =>
-  request.post(api.googleDriveWebAuthStart, { data: payload });
+  request.post(api.googleWebAuthStart('google-drive'), { data: payload });
 
 export const pollGoogleDriveWebAuthResult = (payload: { flow_id: string }) =>
-  request.post(api.googleDriveWebAuthResult, { data: payload });
+  request.post(api.googleWebAuthResult('google-drive'), { data: payload });
+
+// Gmail web auth follows the same pattern as Google Drive, but uses
+// Gmail-specific endpoints and is consumed by the GmailTokenField UI.
+export const startGmailWebAuth = (payload: { credentials: string }) =>
+  request.post(api.googleWebAuthStart('gmail'), { data: payload });
+
+export const pollGmailWebAuthResult = (payload: { flow_id: string }) =>
+  request.post(api.googleWebAuthResult('gmail'), { data: payload });
 
 export default dataSourceService;

web/src/utils/api.ts

@@ -42,8 +42,10 @@ export default {
   dataSourceRebuild: (id: string) => `${api_host}/connector/${id}/rebuild`,
   dataSourceLogs: (id: string) => `${api_host}/connector/${id}/logs`,
   dataSourceDetail: (id: string) => `${api_host}/connector/${id}`,
-  googleDriveWebAuthStart: `${api_host}/connector/google-drive/oauth/web/start`,
+  googleWebAuthStart: (type: 'google-drive' | 'gmail') =>
-  googleDriveWebAuthResult: `${api_host}/connector/google-drive/oauth/web/result`,
+    `${api_host}/connector/google/oauth/web/start?type=${type}`,
+  googleWebAuthResult: (type: 'google-drive' | 'gmail') =>
+    `${api_host}/connector/google/oauth/web/result?type=${type}`,
 
   // plugin
   llm_tools: `${api_host}/plugin/llm_tools`,