Unified Synchronization

Sync your Prompts and Configs across every AI tool.

memory-sync is a high-performance, Rust-first pipeline that synchronizes prompts, rules, and workspace memory into native tool configurations.

Get Started

View on GitHub

ArchitectureRust-first / NAPI-first
Source ModelMDX as source of truth
OutputNative config materialization

Contributors

Built by Real Contributors

Live from GitHub repository metadata and commit co-author records. 4 contributor identities are currently shown here.

View Contributors on GitHub

The Pillars

Explore the Core Sections

Our documentation is divided into seven logical pillars to help you find exactly what you need.

Quick Guide

Decide whether to start with CLI, GUI, or MCP from a single page, then jump to the shortest workable path.

Learn more →CLI

Organized around installation, project setup, the first install run, configuration fields, and the exposed command surface.

Learn more →SDK

Explains the private mixed core, its ownership boundaries, its consumers, and how it relates to the tnmsd crate and CLI.

Learn more →MCP

Covers the stdio server, the exposed tools, workspaceDir semantics, and the integration boundary.

Learn more →GUI

Focuses on the desktop layer, its page structure, and how it works with sdk, the tnmsd crate, and the CLI.

Learn more →Technical Details

Concentrates the architecture boundaries, the install pipeline, the source-of-truth model, and authoring conventions.

Learn more →Design Rationale

Keeps project motivation, design reasoning, and manifesto-style content separate from usage docs.

Learn more →

Capabilities

Built for Modern Workflows

From prompt engineering to automated cleanup, memory-sync provides the tools to keep your AI environment consistent.

Prompt Source

Maintain Input Assets in One Place

Use `.src.mdx` files and structured front matter to maintain prompts, skills, commands, sub-agents, and rules.

Pipeline

Materialize Native Outputs Per Tool

Let plugins declare output targets, cleanup scopes, and generation strategy instead of copying config by hand.

Protection

Control Deletion Boundaries Explicitly

`cleanupProtection` and `outputScopes` define cleanup scope and risk boundaries.

Docs

Document Only What the Repo Actually Does

Pages stay aligned with the current implementation, CLI help, schema, and workflow behavior instead of carrying old narratives forward.

Recommended Path

Getting Up to Speed

Follow this path to quickly establish the correct mental model and start syncing.

01Choose Your Entry Point

Decide whether you are starting from terminal install, the desktop workflow, or MCP integration.

02Confirm Runtime Requirements

Verify Node, Rust, the CLI entrypoint, and the currently supported boundaries first.

03Prepare the Project Structure

Organize source files around the actual aindex and project-config directory conventions.

04Start Maintaining Source Content

Establish the source-of-truth model first, then separate global prompts, workspace prompts, and other input assets.

05Run dry-run and Install

Validate the output list, scope, and cleanup behavior before writing into target tools.