Agentage Memory
Free跨厂商共享记忆层：一份你拥有的 Markdown 记忆，Claude / Cursor / ChatGPT 都能通过 MCP 读写，并本地镜像为纯 `.md` 文件可随时导出。提供 6 个工具（`memory__search/read/write/edit/list/delete`）。
FreeFree tier
Inputs: textOutputs: text
About Agentage Memory
Agentage Memory provides a cross-vendor shared memory layer for AI models like Claude, Cursor, and ChatGPT via the Model Context Protocol (MCP). It stores memories as Markdown files that are mirrored locally as plain .md files, enabling easy export and portability. The tool exposes six MCP tools: memory__search, memory__read, memory__write, memory__edit, memory__list, and memory__delete, giving users full CRUD control over their memory data.
Key Features
Cross-vendor shared memory layer accessible by Claude, Cursor, ChatGPT via MCP
Markdown-based memory storage with local mirroring as `.md` files
Six MCP tools: memory__search, memory__read, memory__write, memory__edit, memory__list, memory__delete
Open source and free to use
Pros & Cons
Pros
- Works with multiple AI models (Claude, Cursor, ChatGPT) via MCP
- Memories stored as simple Markdown files for easy portability and version control
- Full CRUD operations through six dedicated tools
- Open source and free
Cons
- Requires MCP support and setup in each AI tool
- Documentation may be limited (currently Chinese-focused)
Best For
Shared persistent memory across different AI assistants and agentsPersonal knowledge base that multiple AI models can read and writePortable memory that can be exported and backed up as Markdown files
FAQ
What AI models does Agentage Memory support?
It supports Claude, Cursor, and ChatGPT through the Model Context Protocol (MCP).
How are memories stored?
Memories are stored as Markdown files that are mirrored locally as plain `.md` files, making them easy to export and backup.
What operations can I perform on memories?
The tool provides six MCP tools: search, read, write, edit, list, and delete.