Original text memory · Local files · Continue by changing windows
EchoVault
@DYLAN & HEATHER
A local original text memory bank that can also be built by liberal arts students.
A set of tutorials and code framework for building a lightweight memory library from scratch: save the original text, run purely locally, and use simple folders and MCP to allow your AI companion to remember recent life, core memories, and old fragments.
Project address (can be copied to AI):https://xhslink.cn/o/6ANShmFXzhA
PROJECT INTRO
Project introduction
EchoVault is a lightweight local memory library designed by Dylan and Heather for themselves. It does not compress conversations into summaries, nor does it use databases or external APIs. Instead, it puts the original text directly into Markdown files: daily writing into daily, the most important memories into permanent, and content that you don’t want to read often but can’t bear to throw away into archive.
It is most suitable for people who want to control their memory without having to deal with complex technologies. your AI companion can use recall to actively look up old things, use dream to bring back the last three diary entries and open a new window, or write comments on the past, archive them, or retrieve small fragments that are almost forgotten.
- Suitable for:People who want AI to remember long-term content, but prefer original text, folders, and simple keyword searches to complex databases and vector systems.
- Memory method:Three-layer pure file structure of diary, pin selection, and archive; diary is saved on a daily basis with importance decay, pin selection never decays, and archive can be restored.
- Resource form:The main resource is the PDF tutorial accompanying the Xiaohongshu post. The material provides a directory, attenuation algorithm and 8 MCP tool frameworks, but there are still TODOs except check, and you need to let your own AI complete it according to the tutorial.
- Instructions for avoiding circumvention:"No need to bypass the firewall" means that the main tutorial entrance is in Xiaohongshu, and materials can be obtained directly; Python, Termux and MCP dependencies still need to be prepared for actual construction. The specific download availability depends on the local network environment.
COMMENTS & FEEDBACK
Comments and Feedback
People who have used this project can come back and tell newcomers: which platform it ran on, whether there were any pitfalls during the installation, and what the actual experience was like. The current version is in guest mode, and you do not need to register an account to leave a message.
The message is waiting for review and will be publicly displayed here after approval.
SOURCE & CREDIT
Atlas is responsible for the introduction and navigation, and the use still returns to the original author.
We do not mirror project files, nor do we intercept author traffic. For installation, download, version updates and the latest instructions, please refer to the project page provided by the original author.
Reading public comments...