Skip to content

Collections

Collection Management

Collections are logical containers that group related memories. They are automatically created and managed by the system, so manual collection operations are rarely needed.

create_collection

Description: Create a new storage space for memories with specific vector configuration.

Parameters: - collection_name (required): Unique name for the collection - description (optional): Description of the collection's purpose - vector_size (optional): Vector dimensions (default: auto-detected from embedding model) - distance (optional): Distance metric - "Cosine", "Euclidean", or "Dot" (default: "Cosine")

Example: 

curl -X POST https://akashic-mcp.com/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_MCP_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"collections","arguments":{"action":"create","collection_name":"project-notes","description":"Notes and decisions for my project","vector_size":768,"distance":"Cosine"}},"id":1}'

Response: 

{
  "result": "Collection 'project-notes' created successfully",
  "collection_name": "project-notes"
}

Use Cases: - Creating a collection with custom vector configuration - Organizing memories into named collections before storing - Setting up collections for specific projects or contexts

Important Notes: - Collections are usually created automatically when storing memories - Each collection is isolated for security - Vector size should match your embedding model dimensions

list_collections

Description: List all available storage spaces (collections) for the current owner.

Parameters: None

Example: 

curl -X POST https://akashic-mcp.com/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_MCP_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"collections","arguments":{"action":"list"}},"id":1}'

Response: 

[
  {
    "name": "memory-cf17f249-913b-49bf-88b6-f72c7552913c",
    "description": "Default memory collection",
    "vector_size": 768,
    "distance": "Cosine"
  },
  {
    "name": "project-notes",
    "description": "Notes and decisions for my project",
    "vector_size": 768,
    "distance": "Cosine"
  }
]

Use Cases: - Verifying which collections exist - Debugging collection access issues - Viewing collection configurations

Important Notes: - Collections are automatically managed - you rarely need to list them - Only shows collections belonging to the current owner - Collections are transparent - focus on memories, not collection management

delete_collection

Description: Permanently delete a collection and all its memories. This is a destructive operation.

Parameters: - collection_name (required): Name of the collection to delete

Example: 

curl -X POST https://akashic-mcp.com/mcp \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_MCP_TOKEN" \
  -d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"collections","arguments":{"action":"delete","collection_name":"test-collection"}},"id":1}'

Response: 

Collection deletion completed:

✅ Deleted collection: "test-collection"
✅ Deleted 15 memories

Use Cases: - Removing test or temporary collections - Cleaning up outdated collections - Deleting collections with incorrect data

Important Notes: - This operation cannot be undone - Only deletes collections belonging to the current owner - All memories in the collection are permanently deleted