Voice assistant with persistent memory powered by Supermemory and Pipecat.

Quick Start

bun i && bun run setup:backend

# backend/.env
OPENAI_API_KEY=
SUPERMEMORY_API_KEY=

bun run dev:backend
bun run dev

Pipecat Memory Integration

Add persistent memory to your Pipecat voice AI agent. Remember user preferences, past conversations, and context across sessions.

Supermemory integrates with Pipecat, providing long-term memory capabilities for voice AI agents. Your Pipecat applications will remember past conversations and provide personalized responses based on user history.

Installation

pip install supermemory-pipecat

Configuration

from supermemory_pipecat import SupermemoryPipecatService, InputParams

memory = SupermemoryPipecatService(
    api_key=os.getenv("SUPERMEMORY_API_KEY"),
    user_id="user_123",
    params=InputParams(mode="full", search_limit=10),
)

Modes: profile (user facts) | query (search) | full (both)

Pipeline Integration

pipeline = Pipeline([
    transport.input(),
    stt,
    context_aggregator.user(),
    memory,  # Supermemory
    llm,
    tts,
    transport.output(),
    context_aggregator.assistant(),
])

Links

• Supermemory Docs
• Pipecat Docs