## Stack Context
Part 2 of a 3-PR stack moving memory deduplication into the SDKs. See `sdk-dedup/tools-ts` (parent) for the full context and the TypeScript implementation this mirrors.
## What?
Port the normalized, priority-ordered (`static > dynamic > search`) profile deduplication into the Python SDKs.
- Each request injects one **owned memory block that replaces** the prior block rather than accumulating.
- Dedup is **request-local** (no shared state), so it stays correct under concurrency.
Covers OpenAI, Agent Framework (middleware + context provider), Cartesia, and Pipecat.
## Why?
Keeps the Python SDKs at behavioral parity with the TypeScript SDK so all integrations deduplicate memory the same way.
## Testing
- OpenAI: 31 passed, 11 skipped (live)
- Agent Framework: 59 passed
- Cartesia: 8 passed
- Pipecat: 8 passed
🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> **Medium Risk**
> Changes memory formatting and system-prompt injection across multiple SDK integrations; incorrect dedup or replacement could alter LLM context, but there is no auth or data-store risk.
>
> **Overview**
> Ports **normalized cross-source memory deduplication** and **replace-not-append injection** into the Python OpenAI, Agent Framework, Cartesia, and Pipecat packages so they match the TypeScript SDK behavior.
>
> **Deduplication** uses request-local keys: strip optional `[YYYY-MM-DD]` prefixes, normalize whitespace, and compare with `casefold`, with priority **static → dynamic → search**. In **`query` mode**, profile static/dynamic are excluded from dedup input so facts that only appear in search (or overlap profile) are not dropped before formatting.
>
> **Injection** no longer appends memory text every turn. OpenAI and Agent Framework middleware **strip prior owned `<supermemory context="user-memories" readonly>` blocks** and **replace** them once per request while keeping the caller’s system instructions; extra system messages lose stale blocks only. New helpers (`strip`/`replace`/`wrap`) live in each package’s utils.
>
> Tests cover normalized fact variants, query-mode search retention, and stale block replacement.
>
> <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit 42f308b224. Bugbot is set up for automated code reviews on this repo. Configure [here](https://www.cursor.com/dashboard/bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
### TL;DR
Added Python SDK for integrating Supermemory with Cartesia Line voice agents, enabling persistent memory capabilities.
### What changed?
Created a new Python SDK package (`supermemory_cartesia`) that provides:
- `SupermemoryCartesiaAgent` wrapper class that enhances Cartesia Line agents with memory capabilities
- Memory retrieval and storage functionality that integrates with the Supermemory API
- Utility functions for memory formatting, deduplication, and time formatting
- Custom exception classes for error handling
- Comprehensive documentation and type hints
The implementation includes:
- Memory enrichment for user queries
- Automatic storage of conversation history
- Configurable memory retrieval modes (profile, query, full)
- Background processing to avoid blocking the main conversation flow
### How to test?
```python
from supermemory_cartesia import SupermemoryCartesiaAgent
from line.llm_agent import LlmAgent, LlmConfig
import os
# Create base LLM agent
base_agent = LlmAgent(
model="gemini/gemini-2.5-flash-preview-09-2025",
config=LlmConfig(
system_prompt="You are a helpful assistant.",
introduction="Hello!"
)
)
# Wrap with Supermemory
memory_agent = SupermemoryCartesiaAgent(
agent=base_agent,
api_key=os.getenv("SUPERMEMORY_API_KEY"),
user_id="user-123",
)
# Use memory_agent in your Cartesia Line application
```
### Why make this change?
This SDK enables Cartesia Line voice agents to maintain persistent memory across conversations, enhancing user experience by:
1. Providing contextual awareness of past interactions
2. Remembering user preferences and important information
3. Reducing repetition in conversations
4. Creating more personalized and natural voice interactions
The integration is designed to be lightweight and non-blocking, ensuring that memory operations don't impact the responsiveness of voice interactions.
