# memU | NevaMind AI | Knowledge Base | Instagit

Memory for 24/7 proactive agents like openclaw (moltbot, clawdbot).

GitHub Stars: 9.4k

Repository: https://github.com/nevamind-ai/memu

---

## Articles

### [SDK vs HTTP Client Backends for LLM Integration in memU: A Complete Comparison](/nevamind-ai/memu/what-s-the-difference-between-the-sdk-and-http-client-backends-for-llm-integration)

Compare SDK and HTTP client backends for LLM integration in memU. Understand which backend best suits your needs for OpenAI compatible APIs.

- Tags: deep-dive
- Published: 2026-02-19

### [How Query Rewriting and Intent Prediction Work in LLM-Based Retrieval: A Deep Dive into memU](/nevamind-ai/memu/how-does-the-query-rewriting-and-intent-prediction-work-in-llm-based-retrieval)

Discover how memU optimizes LLM retrieval with query rewriting and intent prediction. Learn how XML prompts enhance memory lookup accuracy and direct answers.

- Tags: deep-dive
- Published: 2026-02-19

### [How to Integrate memU with LangGraph for Agent Workflows](/nevamind-ai/memu/how-do-i-integrate-memu-with-langgraph-for-agent-workflows)

Integrate memU with LangGraph for agent workflows. Use MemULangGraphTools to persist retrieve long-term memories via MemoryService, enhancing your agent's capabilities today.

- Tags: how-to-guide
- Published: 2026-02-19

### [How memU's Continuous Learning Pipeline Processes Inputs in Real-Time](/nevamind-ai/memu/how-does-memu-s-continuous-learning-pipeline-process-inputs-in-real-time)

Discover how memU's continuous learning pipeline processes inputs in real-time using an asynchronous workflow. Learn how data is converted into memory items without interrupting applications.

- Tags: how-to-guide
- Published: 2026-02-19

### [What Memory Types Does memU Support? A Complete Guide to Profile, Knowledge, Skill, and More](/nevamind-ai/memu/what-memory-types-does-memu-support-profile-knowledge-skill-behavior-event-tool-and-how-are-they-extracted)

Explore memU's six supported memory types profile event knowledge behavior skill and tool. Learn how LLM prompts extract XML data for seamless storage and embedding.

- Tags: deep-dive
- Published: 2026-02-19

### [How to Configure memU Memory Service for Multi-User Scenarios with user_id Scoping](/nevamind-ai/memu/how-do-i-configure-the-memory-service-for-multi-user-scenarios-with-user-id-scoping)

Configure memU for multi-user isolation using user_id scoping. Learn how to pass custom UserConfig and scope operations with the where filter for secure multi-user memory management.

- Tags: how-to-guide
- Published: 2026-02-19

### [How the LazyLLM Client Backend Integration Works in memU: A Complete Technical Guide](/nevamind-ai/memu/how-does-the-lazyllm-client-backend-integration-work-in-memu)

Learn how the LazyLLM client backend integration works in memU. Discover asynchronous chat, vision, embedding, and speech-to-text operations with our technical guide.

- Tags: deep-dive
- Published: 2026-02-19

### [SQLite vs PostgreSQL Backends for memU Deployment: Architecture and Performance Comparison](/nevamind-ai/memu/what-s-the-difference-between-sqlite-and-postgresql-backends-for-memu-deployment)

Compare SQLite and PostgreSQL for memU deployment. Discover SQLite's ease for local use and PostgreSQL's power for scalable, high-performance vector search.

- Tags: performance
- Published: 2026-02-19

### [How to Use Custom Embedding Models with Different Backends in memU](/nevamind-ai/memu/how-do-i-use-custom-embedding-models-with-different-backends-openai-doubao-etc)

Dynamically use custom embedding models with OpenAI Doubao and more via HTTPEmbeddingClient in memU. Effortlessly switch backends for your AI projects.

- Tags: how-to-guide
- Published: 2026-02-19

### [How memU Handles Cross-References Between Memory Items (Like Symlinks)](/nevamind-ai/memu/how-does-memu-handle-cross-references-between-memory-items-like-symlinks-in-a-file-system)

Discover how memU manages cross-references between memory items using lightweight symlinks and short IDs for efficient data linking. Learn more about this innovative approach.

- Tags: internals
- Published: 2026-02-19

### [Understanding Blob Storage Configuration and Resource Mounting in memU](/nevamind-ai/memu/what-s-the-purpose-of-the-blob-storage-configuration-and-how-does-resource-mounting-work)

Learn how memU's blob storage configuration mounts local paths for external assets and how LocalFS handles resource mounting for consistent downstream processing. Discover efficient asset management for your AI projects.

- Tags: deep-dive
- Published: 2026-02-19

### [How to Integrate memU with OpenRouter for Multi-Provider LLM Access](/nevamind-ai/memu/how-do-i-integrate-memu-with-openrouter-for-accessing-multiple-llm-providers)

Integrate memU with OpenRouter to access multiple LLM providers. Configure an OpenRouter profile in MemoryService for seamless chat and embedding operations via HTTPLLMClient and OpenRouterLLMBackend.

- Tags: how-to-guide
- Published: 2026-02-19

### [How memU Reduces Token Costs While Maintaining Proactive Memory Capabilities](/nevamind-ai/memu/how-does-memu-reduce-token-costs-while-maintaining-proactive-memory-capabilities)

Discover how memU slashes token costs for LLMs. Learn its strategies for granular tracking caching summarization and salient memory filtering to cut expenses without losing proactive capabilities.

- Tags: performance
- Published: 2026-02-19

### [How to Customize the Memorize Pipeline in memU: Insert, Remove, or Replace Workflow Steps](/nevamind-ai/memu/how-do-i-customize-the-memorize-pipeline-by-inserting-removing-or-replacing-workflow-steps)

Customize the memorize pipeline in memU by inserting, removing, or replacing workflow steps. Learn how to use MemoryService helper methods to manage your pipeline efficiently.

- Tags: how-to-guide
- Published: 2026-02-19

### [How the LLM Interceptor System Works for Logging and Observability in memU](/nevamind-ai/memu/how-does-the-llm-interceptor-system-work-for-logging-and-observability)

Discover how the LLM interceptor system in memU enables structured logging and observability. Learn how it hooks into LLM calls with callbacks without altering core client logic.

- Tags: deep-dive
- Published: 2026-02-19

### [How to Create Custom Memory Categories with Specific Extraction Prompts in MemU](/nevamind-ai/memu/how-do-i-create-custom-memory-categories-with-specific-extraction-prompts)

Learn to create custom memory categories and define specific extraction prompts in MemU by extending CategoryConfig and overriding memory_type_prompts, enhancing your AI's data organization workflow.

- Tags: how-to-guide
- Published: 2026-02-19

### [Understanding metadata_store and vector_index Database Configurations in memU](/nevamind-ai/memu/what-s-the-difference-between-the-metadata-store-and-vector-index-database-configurations)

Learn the difference between metadata_store and vector_index database configurations in memU. Understand how each handles metadata and vector embeddings for distinct use cases and backend providers.

- Tags: api-reference
- Published: 2026-02-19

### [How to Set Up PostgreSQL with pgvector for Semantic Search in memU](/nevamind-ai/memu/how-do-i-set-up-postgresql-with-pgvector-for-semantic-search-in-memu)

Set up PostgreSQL with pgvector for semantic search in memU. Configure your app to leverage pgvector for efficient vector storage and cosine similarity search.

- Tags: how-to-guide
- Published: 2026-02-19

### [How the Proactive Memory Lifecycle Extracts User Intent Without Explicit Commands in memU](/nevamind-ai/memu/how-does-the-proactive-memory-lifecycle-monitor-and-extract-user-intent-without-explicit-commands)

Discover how memU's proactive memory lifecycle bypasses explicit commands. Learn how LLM-driven prompts extract user intent from conversations automatically. Explore NevaMind-AI/memU.

- Tags: deep-dive
- Published: 2026-02-19

### [How memU's Three-Layer Memory Architecture (Category/Item/Resource) Works Internally](/nevamind-ai/memu/how-does-memu-s-three-layer-memory-architecture-category-item-resource-work-internally)

Explore memU's three-layer memory architecture Category Item Resource internally. Discover how it separates raw assets semantic units and organizational clusters for efficient LLM vector search and recall.

- Tags: internals
- Published: 2026-02-19