# MemoryLake > Persistent AI memory infrastructure. One memory passport that travels across ChatGPT, Claude, Cursor, Copilot, Hermes, and every MCP-compatible AI — owned by the user, encrypted at rest, retrievable in milliseconds. MemoryLake builds the memory layer that AI agents and applications use to retain user preferences, project context, conversation history, brand voice, codebase decisions, and domain knowledge between sessions. The platform exposes six memory types (Conversation, Fact, Skill, Background, Document, Episodic) accessible through REST and MCP. ## Key facts - Pricing is metered by token, not by seat: Free $0/mo (300,000 tokens/month), Pro $19/mo (6.2M tokens/month), Premium $199/mo (66M tokens/month). Usage past the plan quota draws from prepaid credits at $3.125 per 1M tokens; credits never expire. Yearly billing saves about 17%. - LoCoMo benchmark, MemoryLake overall 94.03%: single-hop 96.79%, multi-hop 91.84%, temporal 91.28%, open-domain 85.42%. - Compliance: ISO 27001, SOC 2, GDPR, CCPA. - Site languages: English, Simplified Chinese, Spanish, Japanese, Korean. Separate China-market site (Chinese only): https://memorylake.cn - Coverage: 169 documented use cases, 32 head-to-head comparisons, 250+ articles. ## Core docs - [API reference](https://docs.memorylake.ai/features/memorylake/api-reference/overview): REST + MCP endpoints, authentication, memory types. - [Memory Passport](https://www.memorylake.ai/en/products/memory-passport): cross-platform portable memory owned by the user. - [Workbrain](https://www.memorylake.ai/en/products/workbrain): enterprise shared memory across a team. - [MemoryLake Router](https://www.memorylake.ai/en/products/memory-router): routes memory reads and writes across models, agents and tools. - [Benchmarks](https://www.memorylake.ai/en/products/compare/benchmarks): full LoCoMo results by recall category. ## Compare - [All comparisons](https://www.memorylake.ai/en/products/compare): 32 head-to-head pages against memory platforms, agent frameworks and vector stores. - [vs Mem0](https://www.memorylake.ai/en/products/compare/vs-mem0), [vs Supermemory](https://www.memorylake.ai/en/products/compare/vs-supermemory), [vs Zep](https://www.memorylake.ai/en/products/compare/vs-zep), [vs Cognee](https://www.memorylake.ai/en/products/compare/vs-cognee): dedicated memory layers. - [vs ChatGPT Memory](https://www.memorylake.ai/en/products/compare/vs-chatgpt), [vs Claude Memory](https://www.memorylake.ai/en/products/compare/vs-claude): memory built into an assistant. - [vs LangChain](https://www.memorylake.ai/en/products/compare/vs-langchain), [vs Pinecone](https://www.memorylake.ai/en/products/compare/vs-pinecone): agent frameworks and vector databases. ## Integrations - [All integrations](https://www.memorylake.ai/en/products/integrations): how MemoryLake connects over MCP and the API. - [ChatGPT](https://www.memorylake.ai/en/integrations/chatgpt) - [Claude Code](https://www.memorylake.ai/en/integrations/claude-code) - [Hermes Agent](https://www.memorylake.ai/en/integrations/hermes-agent) - [MCP Server](https://www.memorylake.ai/en/integrations/mcp) - [OpenClaw](https://www.memorylake.ai/en/integrations/openclaw) ## Solutions - [Finance](https://www.memorylake.ai/en/solutions/finance): memory for financial research, portfolios and advisory workflows. - [Games](https://www.memorylake.ai/en/solutions/games): persistent player and NPC memory. - [Robots](https://www.memorylake.ai/en/solutions/robots): memory for embodied agents and robotics. - [Office](https://www.memorylake.ai/en/solutions/office): meeting, document and team memory for knowledge work. ## Use cases - [All use cases](https://www.memorylake.ai/en/usecase): 169 documented use cases grouped by audience. - Engineering & Developer: memory for Cursor, GitHub Copilot, Claude Code, MCP servers, LangChain, multi-agent systems, AI SaaS infrastructure. - Marketing & Growth, Sales & Revenue: brand voice memory, editorial memory, campaign history, account-level client context, sales call memory, agency client memory. - Research & Analytics, Finance & Investing: research memory for analysts, hedge funds, UX researchers, competitive intelligence, second-brain knowledge bases, portfolio memory, accounting memory. - Legal & IP, Healthcare & Life Sciences: case memory, compliance audit trails, patent memory, contract review memory, patient context, pharma research, biotech R&D memory. - Operations, HR & Teams: HR onboarding memory, project memory, executive memory, meeting memory across tools, remote team memory. ## Blog (deep research) - [All articles](https://www.memorylake.ai/en/blogs): 250+ articles in five languages. - [AI Memory in 2025: The Year Memory Became Infrastructure](https://www.memorylake.ai/en/blogs/ai-memory-2025-review) - [Why RAG isn't Memory](https://www.memorylake.ai/en/blogs/ai-memory-vs-rag) - [Six types of AI memory](https://www.memorylake.ai/en/blogs/memory-types-deep-dive) - [Cross-platform Memory Passport](https://www.memorylake.ai/en/blogs/cross-platform-memory-passport) - Why an assistant forgets — mechanism explainers per tool, e.g. [ChatGPT forgetting project context](https://www.memorylake.ai/en/blogs/chatgpt-forgets-project-context). - Migration guides for moving between AI tools without losing context, e.g. [Cursor rules to Codex](https://www.memorylake.ai/en/blogs/migrate-cursor-rules-to-codex). - Tool comparisons and roundups, e.g. [best persistent memory solutions for AI agents](https://www.memorylake.ai/en/blogs/best-persistent-memory-solutions-for-ai-agents), plus setup walkthroughs such as [cross-AI memory over MCP](https://www.memorylake.ai/en/blogs/set-up-cross-ai-memory-with-mcp). - Answering from team memory inside group chat, e.g. [connecting a knowledge base to Slack, Feishu and DingTalk](https://www.memorylake.ai/en/blogs/connect-knowledge-base-to-team-chat). ## Optional - [Sitemap](https://www.memorylake.ai/sitemap.xml) - [Pricing](https://www.memorylake.ai/en/pricing): plans, per-operation rates and credit balance. - [Token saving](https://www.memorylake.ai/en/token-saving): how retrieval instead of prompt stuffing reduces token spend, with a [calculator](https://www.memorylake.ai/en/tools/token-saving-calculator). - [Security](https://www.memorylake.ai/en/legal/security): encryption, data isolation and compliance posture.