Module mcp 

MCP tool functions. Re-exported under post_cortex::mcp so downstream callers writing custom MCP servers can embed them.

Modules

analysis

Analysis, summaries, insights, and session statistics. Analysis, summaries, insights, and session statistics.

error

Typed error hierarchy for the MCP tool layer. Typed error hierarchy for post-cortex-mcp.

query

Structured and keyword-based queries over session context. Structured and keyword-based queries over session context.

schemas

JSON Schema descriptors for every MCP tool. JSON Schema descriptors for every MCP tool.

search

Semantic and embedding-powered search across sessions. Semantic and embedding-powered search across sessions.

session

Session lifecycle: create, load, checkpoint, list, search, metadata. Session lifecycle: create, load, checkpoint, list, search, and metadata.

update_context

Helpers for recording context updates (single and bulk). MCP-side adapter for context-update writes.

workspace

Workspace CRUD and session-to-workspace membership. Workspace CRUD and session-to-workspace membership.

Structs

ContextUpdateItem

A single context update item accepted by the bulk-update tool.

EntityItem

Wire shape for an entity carried by an MCP update_conversation_context call. entity_type is a lowercase string from the closed set: technology, concept, problem, solution, decision, code_component. Unknown values fall back to concept server-side.

MCPToolResult

Standardised result envelope for every MCP tool.

RelationItem

Wire shape for a relation between two named entities. relation_type is a lowercase string from the closed set: required_by, leads_to, related_to, conflicts_with, depends_on, implements, caused_by, solves. Unknown values cause the request to be rejected with InvalidArgument.

Enums

ContextQuery

Typed query descriptors dispatched by query::query_context.

ContextResponse

Typed response payloads returned by query::query_context.

Error

Errors produced by the MCP tool layer.

Interaction

Typed interaction payloads matching MCP interaction_type values.

Functions

add_session_to_workspace

Add a session to a workspace with a specific role.

bulk_update_conversation_context

Record multiple context updates in a single batch via the canonical service. Items that fail translation or persistence are reported in the response payload — the rest still land, matching the legacy gRPC bulk semantics.

create_session_checkpoint

Create a session checkpoint using the global memory system.

create_session_checkpoint_with_system

Create a session checkpoint using an explicit memory system reference.

create_workspace

Create a new workspace with the given name and description.

delete_workspace

Delete a workspace (sessions are not removed).

enable_embeddings

Check whether the embeddings feature is available and return supported models.

find_related_content

Find content related to a topic within a session (embeddings feature required).

get_all_tool_schemas

Return the full set of MCP tool JSON Schema descriptors.

get_entity_importance_analysis

Analyse entity importance with optional minimum-score and limit filters.

get_entity_network_view

Build a network view of entities and their relationships.

get_key_decisions

Extract a timeline of key architectural and technical decisions.

get_key_insights

Extract the top insights from a session ranked by importance.

get_memory_system

Return the global memory system, lazily initialising it if needed.

get_memory_system_with_config

Create a new memory system from the given configuration.

get_service

Return the cached canonical service, building it from the current memory system if necessary. Both update_conversation_context and bulk_update_conversation_context flow through this — no transport is allowed to bypass the canonical impl.

get_session_statistics

Return detailed statistics about a session’s size and activity.

get_structured_summary

Generate a structured summary of the session with optional filtering.

get_tool_catalog

Return a JSON catalog describing every available MCP tool and its usage.

get_vectorization_stats

Retrieve vectorization statistics for the embedding index (embeddings feature required).

get_workspace

Retrieve workspace details including all associated sessions.

inject_memory_system

Inject a pre-built memory system for daemon mode.

list_sessions

List all sessions via the global memory system.

list_sessions_with_storage

List all persisted sessions using a direct storage reference.

list_workspaces

List all workspaces with session counts.

load_session

Load a session by ID via the global memory system.

load_session_checkpoint

Load a session checkpoint using the global memory system.

load_session_checkpoint_with_system

Load a session from a previously saved checkpoint.

load_session_with_system

Load a session by ID using an explicit memory system reference.

mark_important

Mark a specific context update as important within a session.

query_conversation_context

Execute a typed query against a session via the global memory system.

query_conversation_context_with_system

Execute a typed query against a session using an explicit memory system reference.

remove_session_from_workspace

Remove a session from a workspace.

search_sessions

Search sessions by name or description.

semantic_search

Unified semantic search dispatcher supporting session, workspace, and global scopes.

semantic_search_global

Search across all sessions using AI semantic understanding (embeddings feature required).

semantic_search_session

Search within a single session using AI semantic understanding (embeddings feature required).

update_conversation_context

Record a single context update via the canonical PostCortexService::update_context path.

update_session_metadata

Update the name and/or description of a session.

vectorize_session

Manually trigger embedding vectorization for a session (embeddings feature required).

Type Aliases

McpResult

Crate-level result alias.