Skip to Content
The Makinari API repo now includes an MCP Server — connect your AI models directly. View on GitHub →
MCP ServerToolsMemoriesPOST

Memories

Manage agent memories. Actions: “save” (create new), “list” (search/filter), “update” (modify existing), “delete” (remove).

Input Schema

ParameterTypeDescription
actionstringAction to perform on memories. Enum: save, list, update, delete
contentstringThe content to save or update. Required for “save”, optional for “update”.
keystringOptional categorization key (e.g. “user_preferences”). used in save/update.
memory_idstringThe ID of the memory to update or delete. Required for “update” and “delete” actions.
search_querystringOptional search term to filter memories (searches in content, summary, and key).
typestringOptional memory type filter. Default is “assistant_note”.
limitnumberMaximum number of memories to return. Default 10.
client_idstringOptional. Scope to a specific client.
project_idstringOptional. Scope to a specific project.
task_idstringOptional. Scope to a specific task.

REST Endpoint

Actions map to sub-routes: save/save, list/get, update/update, delete/delete.

POST /api/agents/tools/memories/save POST /api/agents/tools/memories/get POST /api/agents/tools/memories/update POST /api/agents/tools/memories/delete
POST /api/agents/tools/memories/save Authorization: Bearer YOUR_API_KEY Content-Type: application/json { "content": "User prefers formal communication style", "key": "user_preferences", "type": "assistant_note" }

Response:

{ "success": true, "memory": { "id": "...", "key": "user_preferences" } }
Last updated on