usecortex-ai 0.3.6__py3-none-any.whl → 0.5.0__py3-none-any.whl

usecortex_ai/search/client.py

@@ -5,9 +5,11 @@ import typing
 from ..core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
 from ..core.request_options import RequestOptions
 from ..types.bm_25_operator_type import Bm25OperatorType
+from ..types.qn_a_search_response import QnASearchResponse
+from ..types.retrieval_result import RetrievalResult
 from ..types.retrieve_mode import RetrieveMode
-from ..types.retrieve_response import RetrieveResponse
-from ..types.search_chunk import SearchChunk
+from ..types.search_mode import SearchMode
+from ..types.supported_llm_providers import SupportedLlmProviders
 from .raw_client import AsyncRawSearchClient, RawSearchClient
 from .types.alpha import Alpha
 
@@ -30,150 +32,51 @@ class SearchClient:
         """
         return self._raw_client
 
-    def qna(
-        self,
-        *,
-        question: str,
-        session_id: str,
-        tenant_id: str,
-        context_list: typing.Optional[typing.Sequence[str]] = OMIT,
-        search_modes: typing.Optional[typing.Sequence[str]] = OMIT,
-        sub_tenant_id: typing.Optional[str] = OMIT,
-        highlight_chunks: typing.Optional[bool] = OMIT,
-        stream: typing.Optional[bool] = OMIT,
-        search_alpha: typing.Optional[float] = OMIT,
-        recency_bias: typing.Optional[float] = OMIT,
-        ai_generation: typing.Optional[bool] = OMIT,
-        top_n: typing.Optional[int] = OMIT,
-        user_name: typing.Optional[str] = OMIT,
-        user_instructions: typing.Optional[str] = OMIT,
-        multi_step_reasoning: typing.Optional[bool] = OMIT,
-        auto_agent_routing: typing.Optional[bool] = OMIT,
-        metadata: typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]] = OMIT,
-        request_options: typing.Optional[RequestOptions] = None,
-    ) -> typing.Optional[typing.Any]:
-        """
-        Ask a question to your uploaded knowledge base and let Cortex AI answer it.
-
-        Parameters
-        ----------
-        question : str
-            The question to be answered
-
-        session_id : str
-            Unique identifier for the conversation session. Keep it same when the current question refers to a previous answer or question
-
-        tenant_id : str
-            Identifier for the tenant/organization
-
-        context_list : typing.Optional[typing.Sequence[str]]
-            List of context strings to provide additional information
-
-        search_modes : typing.Optional[typing.Sequence[str]]
-            List of search modes to use for finding relevant information
-
-        sub_tenant_id : typing.Optional[str]
-            Identifier for sub-tenant within the tenant
-
-        highlight_chunks : typing.Optional[bool]
-            Whether to return text chunks in the response along with final LLM generated answer
-
-        stream : typing.Optional[bool]
-            Whether to stream the response
-
-        search_alpha : typing.Optional[float]
-            Closer to 0.0 means a exact keyword search will be performed, closer to 1.0 means semantics of the search will be considered. In most cases, you wont have to toggle it yourself.
-
-        recency_bias : typing.Optional[float]
-            Bias towards more recent information (0.0 to 1.0)
-
-        ai_generation : typing.Optional[bool]
-            Whether to use AI for generating responses
-
-        top_n : typing.Optional[int]
-            Number of top results to return
-
-        user_name : typing.Optional[str]
-            Name of the user making the request. This helps LLM to know the user's name if semantics around the username are involved in query. Its generally a good practice to include it possible.
-
-        user_instructions : typing.Optional[str]
-            Custom instructions for the AI response to add to our proprietary prompt. This can be used to provide additional context or instructions for the LLM to follow so that the answers are tailored towards your application style
-
-        multi_step_reasoning : typing.Optional[bool]
-            Enable advanced multi-step reasoning for complex queries. When enabled, the AI will automatically break down complex questions into multiple research steps to provide more comprehensive and accurate answers.
-
-        auto_agent_routing : typing.Optional[bool]
-            Enable intelligent agent routing to automatically select the most suitable AI agent for your specific query type. Different agents are optimized for various use cases like social media, code, conversations, general knowledge, etc.
-
-        metadata : typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]]
-            Additional metadata for the request
-
-        request_options : typing.Optional[RequestOptions]
-            Request-specific configuration.
-
-        Returns
-        -------
-        typing.Optional[typing.Any]
-            Successful Response
-
-        Examples
-        --------
-        from usecortex-ai import CortexAI
-
-        client = CortexAI(token="YOUR_TOKEN", )
-        client.search.qna(question='What is Cortex AI', session_id='chat_session_1234', tenant_id='tenant_1234', )
-        """
-        _response = self._raw_client.qna(
-            question=question,
-            session_id=session_id,
-            tenant_id=tenant_id,
-            context_list=context_list,
-            search_modes=search_modes,
-            sub_tenant_id=sub_tenant_id,
-            highlight_chunks=highlight_chunks,
-            stream=stream,
-            search_alpha=search_alpha,
-            recency_bias=recency_bias,
-            ai_generation=ai_generation,
-            top_n=top_n,
-            user_name=user_name,
-            user_instructions=user_instructions,
-            multi_step_reasoning=multi_step_reasoning,
-            auto_agent_routing=auto_agent_routing,
-            metadata=metadata,
-            request_options=request_options,
-        )
-        return _response.data
-
     def retrieve(
         self,
         *,
-        query: str,
         tenant_id: str,
+        query: str,
         sub_tenant_id: typing.Optional[str] = OMIT,
         max_chunks: typing.Optional[int] = OMIT,
         mode: typing.Optional[RetrieveMode] = OMIT,
         alpha: typing.Optional[Alpha] = OMIT,
         recency_bias: typing.Optional[float] = OMIT,
+        num_related_chunks: typing.Optional[int] = OMIT,
         personalise_search: typing.Optional[bool] = OMIT,
         graph_context: typing.Optional[bool] = OMIT,
         extra_context: typing.Optional[str] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
+        filters: typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
-    ) -> RetrieveResponse:
+    ) -> RetrievalResult:
         """
-        Search for relevant content within your indexed sources.
+        Search for relevant content within your indexed sources or user memories.
+
+            Results are ranked by relevance and can be customized with parameters like
+            result limits, alpha weighting, and recency preferences.
+
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories (uses inferred content)
 
-        This API returns the chunks related to the query you make. We use neural (embedding) search to give you the most relevant chunks.
-        Results are ranked by relevance and can be customized with parameters like result limits and recency preferences.
+            Use `mode` to control retrieval quality:
+            - "fast" (default): Single query, faster response
+            - "accurate": Multi-query generation with reranking, higher quality
+
+            Use `filters` to narrow results by metadata:
+            - Provide key-value pairs matching fields defined in your tenant_metadata_schema
+            - Example: `{"category": "engineering", "priority": "high"}`
+            - Filters are validated against your tenant schema for type safety
 
         Parameters
         ----------
-        query : str
-            Search terms to find relevant content
-
         tenant_id : str
             Unique identifier for the tenant/organization
 
+        query : str
+            Search terms to find relevant content
+
         sub_tenant_id : typing.Optional[str]
             Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
 
@@ -187,7 +90,10 @@ class SearchClient:
             Search ranking algorithm parameter (0.0-1.0 or 'auto')
 
         recency_bias : typing.Optional[float]
-            Preference for newer content (0.0 = no bias, 1.0 = strong recency preference)
+            Preference for newer content (0.0 = no bias, 1.0 =             strong recency preference)
+
+        num_related_chunks : typing.Optional[int]
+            Number of related content chunks to include
 
         personalise_search : typing.Optional[bool]
             Enable personalized search results based on user preferences
@@ -198,12 +104,18 @@ class SearchClient:
         extra_context : typing.Optional[str]
             Additional context provided by the user to guide retrieval
 
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
+
+        filters : typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]]
+            Optional key-value pairs to filter search results by tenant metadata fields. Keys must match fields defined in tenant_metadata_schema during tenant creation. Supports exact match filtering on indexed metadata fields. Example: {'category': 'engineering', 'priority': 'high'}
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
         Returns
         -------
-        RetrieveResponse
+        RetrievalResult
             Successful Response
 
         Examples
@@ -211,19 +123,22 @@ class SearchClient:
         from usecortex-ai import CortexAI
 
         client = CortexAI(token="YOUR_TOKEN", )
-        client.search.retrieve(query='Which mode does user prefer', tenant_id='tenant_1234', )
+        client.search.retrieve(tenant_id='tenant_id', query='query', )
         """
         _response = self._raw_client.retrieve(
-            query=query,
             tenant_id=tenant_id,
+            query=query,
             sub_tenant_id=sub_tenant_id,
             max_chunks=max_chunks,
             mode=mode,
             alpha=alpha,
             recency_bias=recency_bias,
+            num_related_chunks=num_related_chunks,
             personalise_search=personalise_search,
             graph_context=graph_context,
             extra_context=extra_context,
+            search_mode=search_mode,
+            filters=filters,
             request_options=request_options,
         )
         return _response.data
@@ -231,27 +146,31 @@ class SearchClient:
     def full_text_search(
         self,
         *,
-        query: str,
         tenant_id: str,
+        query: str,
         sub_tenant_id: typing.Optional[str] = OMIT,
         operator: typing.Optional[Bm25OperatorType] = OMIT,
         max_chunks: typing.Optional[int] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
-    ) -> typing.List[SearchChunk]:
+    ) -> RetrievalResult:
         """
-        Perform full text search for exact matches within your indexed sources.
+        Perform full text search for exact matches within your indexed sources or memories.
+            Choose between 'OR' and 'AND' operators to control how search terms are combined
+            for precise text matching.
 
-        Use this endpoint to find content chunks using BM25-based text matching with configurable operators.
-        Choose between 'OR' and 'AND' operators to control how search terms are combined for precise text matching.
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories
 
         Parameters
         ----------
-        query : str
-            Search terms to find in your content
-
         tenant_id : str
             Unique identifier for the tenant/organization
 
+        query : str
+            Search terms to find in your content
+
         sub_tenant_id : typing.Optional[str]
             Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
 
@@ -261,12 +180,15 @@ class SearchClient:
         max_chunks : typing.Optional[int]
             Maximum number of results to return
 
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
         Returns
         -------
-        typing.List[SearchChunk]
+        RetrievalResult
             Successful Response
 
         Examples
@@ -274,182 +196,186 @@ class SearchClient:
         from usecortex-ai import CortexAI
 
         client = CortexAI(token="YOUR_TOKEN", )
-        client.search.full_text_search(query='John Smith Jake', tenant_id='tenant_1234', )
+        client.search.full_text_search(tenant_id='tenant_id', query='query', )
         """
         _response = self._raw_client.full_text_search(
-            query=query,
             tenant_id=tenant_id,
+            query=query,
             sub_tenant_id=sub_tenant_id,
             operator=operator,
             max_chunks=max_chunks,
+            search_mode=search_mode,
             request_options=request_options,
         )
         return _response.data
 
-
-class AsyncSearchClient:
-    def __init__(self, *, client_wrapper: AsyncClientWrapper):
-        self._raw_client = AsyncRawSearchClient(client_wrapper=client_wrapper)
-
-    @property
-    def with_raw_response(self) -> AsyncRawSearchClient:
-        """
-        Retrieves a raw implementation of this client that returns raw responses.
-
-        Returns
-        -------
-        AsyncRawSearchClient
-        """
-        return self._raw_client
-
-    async def qna(
+    def qna(
         self,
         *,
-        question: str,
-        session_id: str,
         tenant_id: str,
-        context_list: typing.Optional[typing.Sequence[str]] = OMIT,
-        search_modes: typing.Optional[typing.Sequence[str]] = OMIT,
+        question: str,
         sub_tenant_id: typing.Optional[str] = OMIT,
-        highlight_chunks: typing.Optional[bool] = OMIT,
-        stream: typing.Optional[bool] = OMIT,
-        search_alpha: typing.Optional[float] = OMIT,
-        recency_bias: typing.Optional[float] = OMIT,
-        ai_generation: typing.Optional[bool] = OMIT,
-        top_n: typing.Optional[int] = OMIT,
-        user_name: typing.Optional[str] = OMIT,
-        user_instructions: typing.Optional[str] = OMIT,
-        multi_step_reasoning: typing.Optional[bool] = OMIT,
-        auto_agent_routing: typing.Optional[bool] = OMIT,
-        metadata: typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]] = OMIT,
+        max_chunks: typing.Optional[int] = OMIT,
+        mode: typing.Optional[RetrieveMode] = OMIT,
+        alpha: typing.Optional[float] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
+        include_graph_context: typing.Optional[bool] = OMIT,
+        extra_context: typing.Optional[str] = OMIT,
+        llm_provider: typing.Optional[SupportedLlmProviders] = OMIT,
+        model: typing.Optional[str] = OMIT,
+        temperature: typing.Optional[float] = OMIT,
+        max_tokens: typing.Optional[int] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
-    ) -> typing.Optional[typing.Any]:
+    ) -> QnASearchResponse:
         """
-        Ask a question to your uploaded knowledge base and let Cortex AI answer it.
+        Ask a question and get an AI-generated answer based on your indexed sources or memories.
 
-        Parameters
-        ----------
-        question : str
-            The question to be answered
+            The response includes both the AI answer and the source chunks used to generate it,
+            enabling full transparency and citation capabilities.
 
-        session_id : str
-            Unique identifier for the conversation session. Keep it same when the current question refers to a previous answer or question
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories
 
-        tenant_id : str
-            Identifier for the tenant/organization
+            Use `mode` to control retrieval quality:
+            - "fast" (default): Single query, faster response
+            - "accurate": Multi-query generation with reranking, higher quality
 
-        context_list : typing.Optional[typing.Sequence[str]]
-            List of context strings to provide additional information
+        Parameters
+        ----------
+        tenant_id : str
+            Unique identifier for the tenant/organization
 
-        search_modes : typing.Optional[typing.Sequence[str]]
-            List of search modes to use for finding relevant information
+        question : str
+            The question to answer based on indexed sources
 
         sub_tenant_id : typing.Optional[str]
-            Identifier for sub-tenant within the tenant
-
-        highlight_chunks : typing.Optional[bool]
-            Whether to return text chunks in the response along with final LLM generated answer
+            Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
 
-        stream : typing.Optional[bool]
-            Whether to stream the response
+        max_chunks : typing.Optional[int]
+            Maximum number of context chunks to retrieve
 
-        search_alpha : typing.Optional[float]
-            Closer to 0.0 means a exact keyword search will be performed, closer to 1.0 means semantics of the search will be considered. In most cases, you wont have to toggle it yourself.
+        mode : typing.Optional[RetrieveMode]
+            Retrieval mode: 'fast' for single query, 'accurate' for multi-query with reranking
 
-        recency_bias : typing.Optional[float]
-            Bias towards more recent information (0.0 to 1.0)
+        alpha : typing.Optional[float]
+            Hybrid search alpha (0.0 = sparse/keyword, 1.0 = dense/semantic)
 
-        ai_generation : typing.Optional[bool]
-            Whether to use AI for generating responses
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
 
-        top_n : typing.Optional[int]
-            Number of top results to return
+        include_graph_context : typing.Optional[bool]
+            Whether to include knowledge graph context for enhanced answers
 
-        user_name : typing.Optional[str]
-            Name of the user making the request. This helps LLM to know the user's name if semantics around the username are involved in query. Its generally a good practice to include it possible.
+        extra_context : typing.Optional[str]
+            Additional context to guide retrieval and answer generation
 
-        user_instructions : typing.Optional[str]
-            Custom instructions for the AI response to add to our proprietary prompt. This can be used to provide additional context or instructions for the LLM to follow so that the answers are tailored towards your application style
+        llm_provider : typing.Optional[SupportedLlmProviders]
+            LLM provider for answer generation
 
-        multi_step_reasoning : typing.Optional[bool]
-            Enable advanced multi-step reasoning for complex queries. When enabled, the AI will automatically break down complex questions into multiple research steps to provide more comprehensive and accurate answers.
+        model : typing.Optional[str]
+            Specific model to use (defaults to provider's default model)
 
-        auto_agent_routing : typing.Optional[bool]
-            Enable intelligent agent routing to automatically select the most suitable AI agent for your specific query type. Different agents are optimized for various use cases like social media, code, conversations, general knowledge, etc.
+        temperature : typing.Optional[float]
+            LLM temperature for answer generation (lower = more focused)
 
-        metadata : typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]]
-            Additional metadata for the request
+        max_tokens : typing.Optional[int]
+            Maximum tokens for the generated answer
 
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
         Returns
         -------
-        typing.Optional[typing.Any]
+        QnASearchResponse
             Successful Response
 
         Examples
         --------
-        import asyncio
-
-        from usecortex-ai import AsyncCortexAI
+        from usecortex-ai import CortexAI
 
-        client = AsyncCortexAI(token="YOUR_TOKEN", )
-        async def main() -> None:
-            await client.search.qna(question='What is Cortex AI', session_id='chat_session_1234', tenant_id='tenant_1234', )
-        asyncio.run(main())
+        client = CortexAI(token="YOUR_TOKEN", )
+        client.search.qna(tenant_id='tenant_id', question='question', )
         """
-        _response = await self._raw_client.qna(
-            question=question,
-            session_id=session_id,
+        _response = self._raw_client.qna(
             tenant_id=tenant_id,
-            context_list=context_list,
-            search_modes=search_modes,
+            question=question,
             sub_tenant_id=sub_tenant_id,
-            highlight_chunks=highlight_chunks,
-            stream=stream,
-            search_alpha=search_alpha,
-            recency_bias=recency_bias,
-            ai_generation=ai_generation,
-            top_n=top_n,
-            user_name=user_name,
-            user_instructions=user_instructions,
-            multi_step_reasoning=multi_step_reasoning,
-            auto_agent_routing=auto_agent_routing,
-            metadata=metadata,
+            max_chunks=max_chunks,
+            mode=mode,
+            alpha=alpha,
+            search_mode=search_mode,
+            include_graph_context=include_graph_context,
+            extra_context=extra_context,
+            llm_provider=llm_provider,
+            model=model,
+            temperature=temperature,
+            max_tokens=max_tokens,
             request_options=request_options,
         )
         return _response.data
 
+
+class AsyncSearchClient:
+    def __init__(self, *, client_wrapper: AsyncClientWrapper):
+        self._raw_client = AsyncRawSearchClient(client_wrapper=client_wrapper)
+
+    @property
+    def with_raw_response(self) -> AsyncRawSearchClient:
+        """
+        Retrieves a raw implementation of this client that returns raw responses.
+
+        Returns
+        -------
+        AsyncRawSearchClient
+        """
+        return self._raw_client
+
     async def retrieve(
         self,
         *,
-        query: str,
         tenant_id: str,
+        query: str,
         sub_tenant_id: typing.Optional[str] = OMIT,
         max_chunks: typing.Optional[int] = OMIT,
         mode: typing.Optional[RetrieveMode] = OMIT,
         alpha: typing.Optional[Alpha] = OMIT,
         recency_bias: typing.Optional[float] = OMIT,
+        num_related_chunks: typing.Optional[int] = OMIT,
         personalise_search: typing.Optional[bool] = OMIT,
         graph_context: typing.Optional[bool] = OMIT,
         extra_context: typing.Optional[str] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
+        filters: typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
-    ) -> RetrieveResponse:
+    ) -> RetrievalResult:
         """
-        Search for relevant content within your indexed sources.
+        Search for relevant content within your indexed sources or user memories.
+
+            Results are ranked by relevance and can be customized with parameters like
+            result limits, alpha weighting, and recency preferences.
+
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories (uses inferred content)
+
+            Use `mode` to control retrieval quality:
+            - "fast" (default): Single query, faster response
+            - "accurate": Multi-query generation with reranking, higher quality
 
-        This API returns the chunks related to the query you make. We use neural (embedding) search to give you the most relevant chunks.
-        Results are ranked by relevance and can be customized with parameters like result limits and recency preferences.
+            Use `filters` to narrow results by metadata:
+            - Provide key-value pairs matching fields defined in your tenant_metadata_schema
+            - Example: `{"category": "engineering", "priority": "high"}`
+            - Filters are validated against your tenant schema for type safety
 
         Parameters
         ----------
-        query : str
-            Search terms to find relevant content
-
         tenant_id : str
             Unique identifier for the tenant/organization
 
+        query : str
+            Search terms to find relevant content
+
         sub_tenant_id : typing.Optional[str]
             Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
 
@@ -463,7 +389,10 @@ class AsyncSearchClient:
             Search ranking algorithm parameter (0.0-1.0 or 'auto')
 
         recency_bias : typing.Optional[float]
-            Preference for newer content (0.0 = no bias, 1.0 = strong recency preference)
+            Preference for newer content (0.0 = no bias, 1.0 =             strong recency preference)
+
+        num_related_chunks : typing.Optional[int]
+            Number of related content chunks to include
 
         personalise_search : typing.Optional[bool]
             Enable personalized search results based on user preferences
@@ -474,12 +403,18 @@ class AsyncSearchClient:
         extra_context : typing.Optional[str]
             Additional context provided by the user to guide retrieval
 
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
+
+        filters : typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]]
+            Optional key-value pairs to filter search results by tenant metadata fields. Keys must match fields defined in tenant_metadata_schema during tenant creation. Supports exact match filtering on indexed metadata fields. Example: {'category': 'engineering', 'priority': 'high'}
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
         Returns
         -------
-        RetrieveResponse
+        RetrievalResult
             Successful Response
 
         Examples
@@ -490,20 +425,23 @@ class AsyncSearchClient:
 
         client = AsyncCortexAI(token="YOUR_TOKEN", )
         async def main() -> None:
-            await client.search.retrieve(query='Which mode does user prefer', tenant_id='tenant_1234', )
+            await client.search.retrieve(tenant_id='tenant_id', query='query', )
         asyncio.run(main())
         """
         _response = await self._raw_client.retrieve(
-            query=query,
             tenant_id=tenant_id,
+            query=query,
             sub_tenant_id=sub_tenant_id,
             max_chunks=max_chunks,
             mode=mode,
             alpha=alpha,
             recency_bias=recency_bias,
+            num_related_chunks=num_related_chunks,
             personalise_search=personalise_search,
             graph_context=graph_context,
             extra_context=extra_context,
+            search_mode=search_mode,
+            filters=filters,
             request_options=request_options,
         )
         return _response.data
@@ -511,27 +449,31 @@ class AsyncSearchClient:
     async def full_text_search(
         self,
         *,
-        query: str,
         tenant_id: str,
+        query: str,
         sub_tenant_id: typing.Optional[str] = OMIT,
         operator: typing.Optional[Bm25OperatorType] = OMIT,
         max_chunks: typing.Optional[int] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
-    ) -> typing.List[SearchChunk]:
+    ) -> RetrievalResult:
         """
-        Perform full text search for exact matches within your indexed sources.
+        Perform full text search for exact matches within your indexed sources or memories.
+            Choose between 'OR' and 'AND' operators to control how search terms are combined
+            for precise text matching.
 
-        Use this endpoint to find content chunks using BM25-based text matching with configurable operators.
-        Choose between 'OR' and 'AND' operators to control how search terms are combined for precise text matching.
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories
 
         Parameters
         ----------
-        query : str
-            Search terms to find in your content
-
         tenant_id : str
             Unique identifier for the tenant/organization
 
+        query : str
+            Search terms to find in your content
+
         sub_tenant_id : typing.Optional[str]
             Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
 
@@ -541,12 +483,15 @@ class AsyncSearchClient:
         max_chunks : typing.Optional[int]
             Maximum number of results to return
 
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
+
         request_options : typing.Optional[RequestOptions]
             Request-specific configuration.
 
         Returns
         -------
-        typing.List[SearchChunk]
+        RetrievalResult
             Successful Response
 
         Examples
@@ -557,15 +502,126 @@ class AsyncSearchClient:
 
         client = AsyncCortexAI(token="YOUR_TOKEN", )
         async def main() -> None:
-            await client.search.full_text_search(query='John Smith Jake', tenant_id='tenant_1234', )
+            await client.search.full_text_search(tenant_id='tenant_id', query='query', )
         asyncio.run(main())
         """
         _response = await self._raw_client.full_text_search(
-            query=query,
             tenant_id=tenant_id,
+            query=query,
             sub_tenant_id=sub_tenant_id,
             operator=operator,
             max_chunks=max_chunks,
+            search_mode=search_mode,
+            request_options=request_options,
+        )
+        return _response.data
+
+    async def qna(
+        self,
+        *,
+        tenant_id: str,
+        question: str,
+        sub_tenant_id: typing.Optional[str] = OMIT,
+        max_chunks: typing.Optional[int] = OMIT,
+        mode: typing.Optional[RetrieveMode] = OMIT,
+        alpha: typing.Optional[float] = OMIT,
+        search_mode: typing.Optional[SearchMode] = OMIT,
+        include_graph_context: typing.Optional[bool] = OMIT,
+        extra_context: typing.Optional[str] = OMIT,
+        llm_provider: typing.Optional[SupportedLlmProviders] = OMIT,
+        model: typing.Optional[str] = OMIT,
+        temperature: typing.Optional[float] = OMIT,
+        max_tokens: typing.Optional[int] = OMIT,
+        request_options: typing.Optional[RequestOptions] = None,
+    ) -> QnASearchResponse:
+        """
+        Ask a question and get an AI-generated answer based on your indexed sources or memories.
+
+            The response includes both the AI answer and the source chunks used to generate it,
+            enabling full transparency and citation capabilities.
+
+            Use `search_mode` to specify what to search:
+            - "sources" (default): Search over indexed documents
+            - "memories": Search over user memories
+
+            Use `mode` to control retrieval quality:
+            - "fast" (default): Single query, faster response
+            - "accurate": Multi-query generation with reranking, higher quality
+
+        Parameters
+        ----------
+        tenant_id : str
+            Unique identifier for the tenant/organization
+
+        question : str
+            The question to answer based on indexed sources
+
+        sub_tenant_id : typing.Optional[str]
+            Optional sub-tenant identifier used to organize data within a tenant. If omitted, the default sub-tenant created during tenant setup will be used.
+
+        max_chunks : typing.Optional[int]
+            Maximum number of context chunks to retrieve
+
+        mode : typing.Optional[RetrieveMode]
+            Retrieval mode: 'fast' for single query, 'accurate' for multi-query with reranking
+
+        alpha : typing.Optional[float]
+            Hybrid search alpha (0.0 = sparse/keyword, 1.0 = dense/semantic)
+
+        search_mode : typing.Optional[SearchMode]
+            What to search: 'sources' for documents or 'memories' for user memories
+
+        include_graph_context : typing.Optional[bool]
+            Whether to include knowledge graph context for enhanced answers
+
+        extra_context : typing.Optional[str]
+            Additional context to guide retrieval and answer generation
+
+        llm_provider : typing.Optional[SupportedLlmProviders]
+            LLM provider for answer generation
+
+        model : typing.Optional[str]
+            Specific model to use (defaults to provider's default model)
+
+        temperature : typing.Optional[float]
+            LLM temperature for answer generation (lower = more focused)
+
+        max_tokens : typing.Optional[int]
+            Maximum tokens for the generated answer
+
+        request_options : typing.Optional[RequestOptions]
+            Request-specific configuration.
+
+        Returns
+        -------
+        QnASearchResponse
+            Successful Response
+
+        Examples
+        --------
+        import asyncio
+
+        from usecortex-ai import AsyncCortexAI
+
+        client = AsyncCortexAI(token="YOUR_TOKEN", )
+        async def main() -> None:
+            await client.search.qna(tenant_id='tenant_id', question='question', )
+        asyncio.run(main())
+        """
+        _response = await self._raw_client.qna(
+            tenant_id=tenant_id,
+            question=question,
+            sub_tenant_id=sub_tenant_id,
+            max_chunks=max_chunks,
+            mode=mode,
+            alpha=alpha,
+            search_mode=search_mode,
+            include_graph_context=include_graph_context,
+            extra_context=extra_context,
+            llm_provider=llm_provider,
+            model=model,
+            temperature=temperature,
+            max_tokens=max_tokens,
             request_options=request_options,
         )
         return _response.data