bastra-recall

What it does

Bastra-Recall is a persistent memory solution designed for AI coding assistants and language models. It provides a unified, local-first approach to storing and retrieving contextual information across multiple AI tools including Claude Code, Cursor, ChatGPT, and any MCP (Model Context Protocol) client. The tool maintains a Markdown vault that centralizes all memories and context, eliminating the need to re-explain context across different AI applications.

How to set up

Installation involves cloning the repository and configuring it as an MCP server or integrating it with your preferred AI tool. Users create a local Markdown vault directory and configure the connection string in their AI tool's settings. The setup requires basic TypeScript knowledge for custom configurations but provides sensible defaults for immediate use.

Key features

• Local-first architecture with no cloud dependency
• Shared Markdown vault accessible across all connected AI tools
• MCP (Model Context Protocol) compliance for broad compatibility
• Persistent memory that survives across sessions
• Obsidian-compatible format for knowledge management integration
• Zero vendor lock-in with open Markdown storage format