memorer 0.5.1

Semantic memory for conversational AI - remember, recall, forget. Sub-100ms retrieval with emotional ranking and graph reasoning.

Memorer Python SDK

Memory for AI agents. Remember conversations, recall context, forget when needed.

pip install memorer

Usage

from memorer import Memorer

client = Memorer(api_key="mem_sk_...")
user = client.for_user("user-123")

# During conversation, store what matters
user.remember("User mentioned they're allergic to shellfish")
user.remember("Has a meeting with Sarah next Friday at 2pm")
user.remember("Just adopted a dog named Biscuit")

# Before responding, recall relevant context
results = user.recall("any dietary restrictions?")
print(results.context)
# → "User is allergic to shellfish"

results = user.recall("what's on their schedule?")
print(results.context)
# → "Meeting with Sarah on Friday at 2pm"

Conversations

Track conversation history (short-term memory) combined with semantic recall (long-term memory):

# Start or continue a conversation
conv = user.conversation("session-123")  # existing session
# or
conv = user.conversation()  # create new

# Add messages (auto-extracts memories)
conv.add("user", "I just moved to Seattle and love coffee")
conv.add("assistant", "Welcome to Seattle! Great city for coffee lovers.")

# Query with conversation context + long-term memories
result = conv.recall("where does the user live")
print(result.context)  # Combined context ready for LLM
# → Recent conversation + "User lives in Seattle" (extracted memory)

# Get recent messages
messages = conv.messages(limit=20)

Graph Reasoning

Connect memories across conversations:

results = user.recall(
    "gift ideas for them",
    use_graph_reasoning=True,
)
# → Connects: new dog + mentioned liking outdoors → dog hiking gear

Resources

entities = user.entities.list()
memories = user.memories.list()
conversations = user.conversations.list()
communities = client.graph.communities()

Links

• Docs
• API Reference