memory-ferry
Connects GitHub Copilot CLI agents to a shared long-term memory and reusable Skills layer.
Uses a local Ollama instance for optional semantic recall, blending embedding similarity with text scoring in memory and context searches.
Connects OpenAI Codex agents to a shared long-term memory and reusable Skills layer.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@memory-ferryremember that I prefer concise answers in all responses"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Memory Ferry
Memory Ferry · 记忆摆渡 is a local-first, cross-device long-term memory and reusable-Skill layer for AI agents. Put an ordinary Vault folder in iCloud Drive, OneDrive, Dropbox, Syncthing, or a mounted NAS, then connect Codex, Claude Code, Cursor, OpenCode, DeepSeek Harness, Gemini CLI, GitHub Copilot CLI, and MiniMax Code to the same memory.
中文说明 · English user guide · 中文使用手册
Version
0.7.0-beta.1does not provide end-to-end encryption. Never store passwords, API keys, tokens, private keys, recovery codes, or other secrets.
Why it exists
Cross-computer: sync ordinary files without a hosted control plane or database daemon.
Cross-agent: expose one MCP and CLI contract instead of locking memory to a model vendor.
Offline-friendly: each write is an independent immutable JSON event that syncs later.
Conflict-tolerant: concurrent additions never overwrite one another; updates converge deterministically.
Reusable Skills: validate one portable source, render optional Agent overlays, and distribute target-specific packages by content hash on macOS and Windows.
Low cost: reuse storage and sync software you already operate.
Related MCP server: SharedBrain
Supported agents
Agent | MCP | Shared Skills |
OpenAI Codex | ✓ | ✓ |
Claude Code | ✓ | ✓ |
Cursor | ✓ | ✓ |
OpenCode | ✓ | ✓ |
DeepSeek Harness | ✓ | ✓ |
Gemini CLI | ✓ | ✓ |
GitHub Copilot CLI | ✓ | ✓ |
MiniMax Code | ✓ | ✓ |
Quick start
Requires Node.js 20 or newer:
npm install -g memory-ferry@beta
memory-ferry init "/path/to/your/synced/MemoryFerryVault"
memory-ferry agents install all
memory-ferry remember "Prefer concise answers" --kind preference --scope global
memory-ferry recall "answer style"
memory-ferry uiDuring the beta period, use the explicit @beta tag. Upgrade or uninstall with:
npm install -g memory-ferry@beta
npm uninstall -g memory-ferryThe dashboard is loopback-only, opens in your browser, supports English and Chinese, manages multiple Vaults, connects or disconnects agents, reports Skill compatibility, migrates Vault contents safely, and reconciles Skills every five minutes while running.
Portable and Agent-adapted Skills
Every Skill starts with one portable source at <Vault>/skills/<skill-name>/SKILL.md. Memory Ferry validates the open Agent Skills frontmatter before distribution. Run:
memory-ferry skills inspect
memory-ferry skills sync allFor product-specific behavior, add an optional memory-ferry.yaml manifest and small adapters/<agent>/ overlays. An overlay can merge frontmatter.yaml, append APPEND.md, and provide extra files such as Codex agents/openai.yaml; it may not replace the common SKILL.md. Memory Ferry renders and hashes each destination separately, groups Agents that share a discovery directory, blocks conflicting overlays, rejects non-portable symlinks, and preserves unrelated user-owned Skills. See the Skill compatibility guide.
Windows portable app
The Windows x64 release is a portable MemoryFerry.exe: no installer and no separate Node.js installation are required. Put the EXE in a permanent folder and double-click it. It starts without a console window and opens the local dashboard. Use the bundled MemoryFerry-CLI.cmd when you need command-line output.
The beta is not commercially code-signed, so Windows may show an unknown-publisher warning. Verify the included SHA-256 file and keep SmartScreen enabled.
Storage model
MemoryFerryVault/
├── memory-ferry.json
├── memory/events/<device-id>/<date>/<event-id>.json
└── skills/<skill-name>/...Device identity, clocks, event caches, and optional Ollama embedding indexes stay outside the synchronized folder. The Vault contains no live SQLite database, WAL, or lock file.
Upgrade compatibility
Memory Ferry recognizes existing Agent Vault Sync installations:
agent-vault.jsonis accepted and a newmemory-ferry.jsoncompanion manifest is created without deleting the old file.~/.agent-vault-syncstate is migrated to~/.memory-ferrywhen the new state does not exist.AGENT_VAULT_*environment variables remain fallbacks for theirMEMORY_FERRY_*replacements.Reconnecting an agent removes the old managed
agent-vaultentry and writesmemory-ferry.The legacy
agent-vaultCLI alias remains available during the beta migration period.
Back up important Vaults before upgrading and allow your sync client to finish before opening the same Vault on another computer.
Optional semantic recall
If local Ollama is available, recall and context blend semantic similarity with Unicode/CJK text scoring. Set MEMORY_FERRY_EMBED=off or pass --no-semantic to use text-only search. Embedding indexes remain local and are repaired before the first query returns after cloud sync or migration.
Development
npm ci
npm run test:all
npm audit --omit=dev
npm pack --dry-runWindows releases are built and smoke-tested in GitHub Actions. See CONTRIBUTING.md, SECURITY.md, and architecture.md.
License
MIT. Standalone releases also include Node.js and third-party license notices.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Persistent memory and knowledge management for AI agents with semantic search and 50+ tools.
Persistent memory for AI agents. Semantic search, memory graph, W3C DID identity.
Persistent memory for AI agents. Search, store, and recall across sessions.
Persistent memory for AI agents. Search and store durable facts, preferences and decisions.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceProvides persistent AI agent memory using a local vector database for long-term semantic storage and short-term session scratchpads. It enables low-latency memory operations including search, storage, and bulk management without external cloud dependencies.
- AlicenseNot gradedqualityBmaintenanceLocal-first, multi-user shared memory for AI agents with semantic search, offline support, and team synchronization.MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to maintain persistent, local memory with retrieval-augmented search, knowledge graphs, and context surfacing, without any cloud dependencies.135MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents with long-term memory and retrieval-augmented generation (RAG) capabilities, allowing them to recall past conversations, search local files, and learn user preferences.MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/iced7/memory-ferry'
If you have feedback or need assistance with the MCP directory API, please join our Discord server