Biological Memory Consolidation for Cognee

🌐 Website | 🎮 Demo

Nexyn Core is a lightweight Python library that adds human-like memory physics to your AI applications. Built specifically to run alongside Cognee, it solves a major problem in AI memory systems: context bloat.

Instead of treating every piece of information equally, Nexyn automatically scores how important a memory is (Valence), applies a decay rate, and forgets irrelevant data over time—just like a human brain.

• Infinite Context Clutter: AI agents quickly drown in their own logs. Nexyn ensures only the most important, reinforced memories survive long-term.
• Flat Memory Structures: Nexyn categorizes data dynamically. A passing thought decays in hours; a core user instruction (like "I am allergic to peanuts") becomes a permanent instinct.

1. Sensory Buffer: Intercepts data before it enters your vector database.
2. Evaluator: Uses NVIDIA NIM to quickly score the emotional/logical weight (Valence).
3. Retrieval & Rehearsal: Every time you search for a memory, Nexyn resets its decay timer (rehearsal), naturally surfacing things you think about often.
4. Consolidation: A background process sweeps your database and permanently deletes memories that have decayed to zero.

# Using pip
pip install nexyn-core

Nexyn is designed to be completely invisible and is 100% compatible with both Cognee Cloud and Cognee Open Source (Local).

Cloud vs Local Multi-Tenancy: If you are using Cognee Cloud, pass tenant_id and user_id to nexyn.inject() to isolate your users' memory graphs. If you are running Cognee Open Source locally, leave these parameters blank in Nexyn, and instead pass dataset_name="user_123" directly into your cognee.add() calls to partition your data natively!

You simply initialize it once, and it automatically intercepts your native Cognee calls to apply biological physics. You don't need to learn a new API.

import asyncio
import cognee
import nexyn

async def main():
    # 1. Initialize Nexyn's cognitive layer
    # For Cognee Cloud: Pass cognee_url and cognee_api_key
    # For Cognee Open Source (Local): Leave cognee_url empty
    await nexyn.inject(
        nim_api_key="nvapi-your-key-here", 
        cognee_api_key="your_cognee_api_key", # Leave empty for local open source
        cognee_url="https://api.cognee.ai",   # Leave empty for local open source
        tenant_id="default",                  # Leave empty for local open source
        user_id="user_123"                    # Leave empty for local open source
    )

    # 2. Add memories normally (Nexyn automatically scores Valence)
    await cognee.add("Doug is the groom. The wedding is Sunday.")
    
    # 3. Compile the Cognee knowledge graph
    await cognee.cognify()
    
    # 4. Search triggers decay physics and memory rehearsal
    results = await cognee.search("Where is Doug?")
    
    # 5. Fast-forward time to prune dead memories
    await nexyn.sweep()

if __name__ == "__main__":
    asyncio.run(main())

Want to see the biological pipeline in action? 👉 Live Demo

Want to dive deeper into the theory behind Nexyn? Check out our official Medium article: 👉 Nexyn: Teaching AI to forget, so it can actually remember.

Distributed under the MIT License.