Mem0

Agent memory management

Mem0 is a powerful memory management system designed specifically for AI agents. It provides a persistent, searchable memory store that enables agents to remember past interactions, learn from experiences, and maintain context across conversations and workflow executions.

With Mem0, you can:

  • Store agent memories: Save conversation history, user preferences, and important context
  • Retrieve relevant information: Use semantic search to find the most relevant past interactions
  • Build context-aware agents: Enable your agents to reference past conversations and maintain continuity
  • Personalize interactions: Tailor responses based on user history and preferences
  • Implement long-term memory: Create agents that learn and adapt over time
  • Scale memory management: Handle memory needs for multiple users and complex workflows

In Sim, the Mem0 integration enables your agents to maintain persistent memory across workflow executions. This allows for more natural, context-aware interactions where agents can recall past conversations, remember user preferences, and build upon previous interactions. By connecting Sim with Mem0, you can create agents that feel more human-like in their ability to remember and learn from past experiences. The integration supports adding new memories, searching existing memories semantically, and retrieving specific memory records. This memory management capability is essential for building sophisticated agents that can maintain context over time, personalize interactions based on user history, and continuously improve their performance through accumulated knowledge.

Usage Instructions

Integrate Mem0 into the workflow. Can add, search, and retrieve memories.

Tools

mem0_add_memories

Add memories to Mem0 for persistent storage and retrieval

Input

ParameterTypeRequiredDescription
userIdstringYesUser ID associated with the memory (e.g., "user_123", "alice@example.com")
messagesjsonYesArray of message objects with role and content (e.g., [{"role": "user", "content": "Hello"}])
apiKeystringYesYour Mem0 API key

Output

ParameterTypeDescription
idsarrayArray of memory IDs that were created
memoriesarrayArray of memory objects that were created
idstringUnique identifier for the memory
memorystringThe content of the memory
eventstringEvent type indicating operation performed (ADD, UPDATE, DELETE, NOOP)
metadatajsonCustom metadata associated with the memory

mem0_search_memories

Search for memories in Mem0 using semantic search

Input

ParameterTypeRequiredDescription
userIdstringYesUser ID to search memories for (e.g., "user_123", "alice@example.com")
querystringYesSearch query to find relevant memories (e.g., "What are my favorite foods?")
limitnumberNoMaximum number of results to return (e.g., 10, 50, 100)
apiKeystringYesYour Mem0 API key

Output

ParameterTypeDescription
searchResultsarrayArray of search results with memory data and similarity scores
idstringUnique identifier for the memory
memorystringThe content of the memory
user_idstringUser ID associated with this memory
agent_idstringAgent ID associated with this memory
app_idstringApp ID associated with this memory
run_idstringRun/session ID associated with this memory
hashstringHash of the memory content
metadatajsonCustom metadata associated with the memory
categoriesjsonAuto-assigned categories for the memory
created_atstringISO 8601 timestamp when the memory was created
updated_atstringISO 8601 timestamp when the memory was last updated
scorenumberSimilarity score from vector search
idsarrayArray of memory IDs found in the search results

mem0_get_memories

Retrieve memories from Mem0 by ID or filter criteria

Input

ParameterTypeRequiredDescription
userIdstringYesUser ID to retrieve memories for (e.g., "user_123", "alice@example.com")
memoryIdstringNoSpecific memory ID to retrieve (e.g., "mem_abc123")
startDatestringNoStart date for filtering by created_at (e.g., "2024-01-15")
endDatestringNoEnd date for filtering by created_at (e.g., "2024-12-31")
limitnumberNoMaximum number of results to return (e.g., 10, 50, 100)
apiKeystringYesYour Mem0 API key

Output

ParameterTypeDescription
memoriesarrayArray of retrieved memory objects
idstringUnique identifier for the memory
memorystringThe content of the memory
user_idstringUser ID associated with this memory
agent_idstringAgent ID associated with this memory
app_idstringApp ID associated with this memory
run_idstringRun/session ID associated with this memory
hashstringHash of the memory content
metadatajsonCustom metadata associated with the memory
categoriesjsonAuto-assigned categories for the memory
created_atstringISO 8601 timestamp when the memory was created
updated_atstringISO 8601 timestamp when the memory was last updated
ownerstringOwner of the memory
organizationstringOrganization associated with the memory
immutablebooleanWhether the memory can be modified
expiration_datestringExpiration date after which memory is not retrieved
idsarrayArray of memory IDs that were retrieved

On this page

Start building today
Trusted by over 60,000 builders.
Build Agentic workflows visually on a drag-and-drop canvas or with natural language.
Get started