Skip to main content
Glama

create_memory_block

Create structured memory blocks with labels like persona, human, or system for organizing information in the Letta system. Link blocks to agents or update content as needed.

Instructions

Create a new memory block in the Letta system. Common labels: "persona", "human", "system". Use attach_memory_block to link to agents, or update_memory_block to modify later.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
nameYesName of the memory block
labelYesLabel for the memory block (e.g., "persona", "human", "system")
valueYesContent of the memory block
agent_idNoOptional agent ID to create the block for a specific agent
metadataNoOptional metadata for the memory block

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
idYesUnique identifier of the created memory block
nameYes
labelYes
valueNo
metadataNo

Implementation Reference

  • Main handler function for 'create_memory_block' tool. Validates input args (name, label, value), creates memory block via POST /blocks API, optionally attaches to agent if agent_id provided, returns structured content with block_id and details.
    export async function handleCreateMemoryBlock(server, args) {
        try {
            // Validate arguments
            if (!args.name || typeof args.name !== 'string') {
                throw new Error('Missing required argument: name (must be a string)');
            }
            if (!args.label || typeof args.label !== 'string') {
                throw new Error('Missing required argument: label (must be a string)');
            }
            if (!args.value || typeof args.value !== 'string') {
                throw new Error('Missing required argument: value (must be a string)');
            }
    
            // Headers for API requests
            const headers = server.getApiHeaders();
    
            // If agent_id is provided, set the user_id header
            if (args.agent_id) {
                headers['user_id'] = args.agent_id;
            }
    
            // Prepare metadata
            const metadata = args.metadata || {
                type: args.label,
                version: '1.0',
                last_updated: new Date().toISOString(),
            };
    
            // Prepare block data
            const blockData = {
                name: args.name,
                label: args.label,
                value: args.value,
                metadata: metadata,
            };
    
            // Create the memory block
            logger.info(`Creating memory block "${args.name}" with label "${args.label}"...`);
            const createResponse = await server.api.post('/blocks', blockData, { headers });
            const blockId = createResponse.data.id;
    
            // If agent_id is provided, attach the block to the agent
            if (args.agent_id) {
                const attachUrl = `/agents/${args.agent_id}/core-memory/blocks/attach/${blockId}`;
                await server.api.patch(attachUrl, {}, { headers });
    
                // Get agent info
                const agentInfoResponse = await server.api.get(`/agents/${args.agent_id}`, { headers });
                const agentName = agentInfoResponse.data.name || 'Unknown';
    
                return {
                    content: [
                        {
                            type: 'text',
                            text: JSON.stringify({
                                block_id: blockId,
                                name: args.name,
                                label: args.label,
                                agent_id: args.agent_id,
                                agent_name: agentName,
                            }),
                        },
                    ],
                };
            } else {
                // Just return the created block info
                return {
                    content: [
                        {
                            type: 'text',
                            text: JSON.stringify({
                                block_id: blockId,
                                name: args.name,
                                label: args.label,
                            }),
                        },
                    ],
                };
            }
        } catch (error) {
            server.createErrorResponse(error);
        }
    }
  • Tool definition object with name, description, and inputSchema defining required parameters (name, label, value) and optional (agent_id, metadata).
    export const createMemoryBlockToolDefinition = {
        name: 'create_memory_block',
        description:
            'Create a new memory block in the Letta system. Common labels: "persona", "human", "system". Use attach_memory_block to link to agents, or update_memory_block to modify later.',
        inputSchema: {
            type: 'object',
            properties: {
                name: {
                    type: 'string',
                    description: 'Name of the memory block',
                },
                label: {
                    type: 'string',
                    description: 'Label for the memory block (e.g., "persona", "human", "system")',
                },
                value: {
                    type: 'string',
                    description: 'Content of the memory block',
                },
                agent_id: {
                    type: 'string',
                    description: 'Optional agent ID to create the block for a specific agent',
                },
                metadata: {
                    type: 'object',
                    description: 'Optional metadata for the memory block',
                },
            },
            required: ['name', 'label', 'value'],
        },
    };
  • Import of the handler function and tool definition from the implementation file.
        handleCreateMemoryBlock,
        createMemoryBlockToolDefinition,
    } from './memory/create-memory-block.js';
  • Registration/dispatch of the 'create_memory_block' tool call to the specific handler in the central switch statement within registerToolHandlers.
    case 'create_memory_block':
        return handleCreateMemoryBlock(server, request.params.arguments);
  • Output schema for 'create_memory_block' tool responses, defining expected structure (id, name, label, etc.) for structured outputs.
    create_memory_block: {
        type: 'object',
        properties: {
            id: { type: 'string', description: 'Unique identifier of the created memory block' },
            name: { type: 'string' },
            label: { type: 'string' },
            value: { type: 'string' },
            metadata: { type: 'object' },
        },
        required: ['id', 'name', 'label'],
    },
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only provide a title, so the description carries the full burden of behavioral disclosure. It states this is a creation operation (implying mutation) and mentions common labels, but doesn't address permissions, error conditions, rate limits, or what happens when creating duplicate blocks. It adds some context but lacks comprehensive behavioral details.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is perfectly concise with two sentences that each earn their place: the first states the core purpose, the second provides usage guidance and sibling differentiation. No wasted words, and information is front-loaded appropriately.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given that an output schema exists (so return values are documented elsewhere), the description provides good context for a creation tool. It covers purpose, usage guidelines, and parameter hints, though could benefit from more behavioral details about the creation operation's implications in the system.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the schema already documents all 5 parameters thoroughly. The description adds minimal value beyond the schema by mentioning common label examples ('persona', 'human', 'system'), which slightly enhances understanding of the 'label' parameter. This meets the baseline for high schema coverage.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the specific action ('Create a new memory block') and resource ('in the Letta system'), distinguishing it from siblings like 'update_memory_block' (modification) and 'attach_memory_block' (linking). It provides concrete examples of common labels ('persona', 'human', 'system') to clarify the tool's domain.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly provides when-to-use guidance by naming alternatives: 'Use attach_memory_block to link to agents, or update_memory_block to modify later.' This clearly distinguishes this tool's role from sibling tools and provides context for when each should be used.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/oculairmedia/Letta-MCP-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server