Stores a new memory with embedding vector for semantic search and retrieval. The content will be automatically vectorized using embeddings for semantic search capabilities. This endpoint returns immediately and processes the memory in the background.
JWT Authorization header using the Bearer scheme. Example: "Authorization: Bearer {token}"
Unique identifier for the user storing the memory
1"user_123"
The actual content/text to be stored. This can be either a string or a JSON object. The JSON object format is: [{ "role": "user", "content": "string" }, { "role": "assistant", "content": "string" }]
1 - 30000"[{ \"role\": \"user\", \"content\": \"I prefer dark mode\" }, { \"role\": \"assistant\", \"content\": \"I will set the dark mode\" }]"
User consent flag - must be true to store the memory (required for compliance)
true
Project identifier to associate the memory with (optional)
"project_abc"
Optional tags to categorize and filter memories
["preferences", "ui", "notifications"]Optional JSON Metadata to categorize and filter memories
{
"timestamp": "2025-08-02T14:30:00Z",
"category": "user-preferences",
"source": "settings-page"
}Accepted