Awesome AI Memory Systems

         

A curated list of tools, frameworks, and resources for AI memory systems — mechanisms that enable AI applications to store, retrieve, and evolve context over time.

Contents

• Memory Frameworks
• Vector Databases
• Knowledge Stores & Retrieval
• Context Management
• Personalization & User Memory
• Evaluation & Observability
• Research & Concepts

Memory Frameworks

Frameworks for building structured memory layers in AI systems.

• LangChain Memory — Memory abstractions for storing and retrieving conversational context.
• LangGraph — Stateful agent workflows with persistent memory.
• MemGPT — System for managing long-term context using external memory.
• LlamaIndex — Framework for connecting LLMs to external data and memory.
• Semantic Kernel — Memory and planning capabilities for AI applications.

Vector Databases

Databases optimized for storing embeddings and enabling semantic search.

• Pinecone — Managed vector database for semantic search and retrieval.
• Weaviate — Open-source vector database with built-in ML capabilities.
• Milvus — High-performance vector database for large-scale similarity search.
• Qdrant — Vector database with filtering and metadata support.
• Chroma — Lightweight embedding database for AI applications.

Knowledge Stores & Retrieval

Systems for structuring and retrieving long-term knowledge.

• Elasticsearch — Search engine for structured and unstructured data.
• Typesense — Fast, open-source search engine for instant search experiences.
• Meilisearch — Lightweight search engine with typo tolerance and speed.
• Haystack — Framework for building retrieval pipelines.
• FAISS — Library for efficient similarity search and clustering.

Context Management

Tools and approaches for managing prompt context and memory limits.

• tiktoken — Tokenizer for managing input size and context windows.
• Guidance — Framework for controlling LLM outputs and context.
• Guardrails — Framework for enforcing structure and constraints in outputs.
• PromptLayer — Tool for tracking, managing, and versioning prompts.

Personalization & User Memory

Systems focused on storing user-specific context and preferences.

• Mem0 — Lightweight memory layer for personalization in AI agents.
• Zep — Long-term memory store for conversational AI applications.
• Rewind — Personal memory system capturing user activity and context.
• Supermemory — Personal AI memory infrastructure.

Evaluation & Observability

Tools for monitoring memory quality and retrieval effectiveness.

• LangSmith — Observability and debugging for LLM pipelines and memory usage.
• Ragas — Evaluation toolkit for retrieval and memory systems.
• Promptfoo — Testing framework for prompt and memory performance.
• DeepEval — Evaluation framework for LLM systems.

Research & Concepts

Key ideas and research shaping AI memory systems.

• Retrieval-Augmented Generation (RAG) — Combining external knowledge retrieval with generation.
• Long-term memory architectures — Systems that persist and evolve context across sessions.
• Episodic vs semantic memory — Distinction between event-based and knowledge-based storage.
• Context compression — Techniques for summarizing and retaining relevant information.
• Memory retrieval strategies — Ranking, filtering, and relevance scoring for context recall.

Related Awesome Lists

• Awesome AI — General tools, frameworks, and resources for artificial intelligence.
• Awesome AI Agents — Frameworks and tools for building autonomous AI agents.
• Awesome RAG — Tools and frameworks for retrieval-augmented generation.
• Awesome LLMOps — Infrastructure and operational practices for LLM systems.

Contribute

Contributions are welcome. Please ensure your submission fully follows the requirements outlined in CONTRIBUTING.md, including formatting, scope alignment, and category placement.

Pull requests that do not adhere to the contribution guidelines may be closed.

License