project-sync
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., "@project-syncLoad project client-alpha and brief me on where we left off."
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.
CloudGentic Project Sync
No LLM. No API keys. No vector database. One SQLite file on your server, and every AI you use picks up where the last one left off.
A self-hosted Remote MCP server for cross-model session handoff. Start a project in ChatGPT, continue in Claude, finish anywhere: one shared project log, files with version history, and full-text search, all on hardware you control.
There are plenty of AI "memory" servers. This is not a memory database. It is a handoff log: each session ends with a summary and next steps, so the next AI (any AI) starts with the state of the project instead of six months of transcript.
Quick start (Docker)
git clone https://github.com/cloudgentic/project-sync && cd project-sync
docker compose up -d
docker compose exec sync node dist/cli.js token create chatgpt
# copy the token: it is shown exactly oncePut it behind HTTPS (required before exposing to the internet). Minimal Caddy example:
mcp.yourdomain.com {
reverse_proxy 127.0.0.1:3210
}Note: the compose file publishes host port 3210 (container port 3000) to stay clear of the common dev-server port. Change the mapping in docker-compose.yml if you prefer another port; behind Traefik or another container-aware proxy, drop the ports section and route to the container directly.
Related MCP server: stafett
Connect your AI clients (one command)
On your PC, not the server:
npx @cloudgentic/project-sync setupA local page opens (or add --cli for the terminal). Enter your server URL and the admin key the server printed at first start, pick a client, press Connect. The assistant mints a token, appends project-sync to your client config (your other MCP servers are never touched; a backup is saved first), restarts the client, and verifies the connection with a real handshake. Built screen-reader-first: every step is announced, nothing needs a mouse.
Works today with Claude Desktop (Windows, macOS, Linux) and Claude Code. Claude web/mobile and ChatGPT custom connectors need OAuth, arriving in a later release. Manage tokens later with npx @cloudgentic/project-sync setup manage --server <url> --admin-key <key>.
Prefer to do it by hand? See docs/ for the manual config blocks and the paste-once instructions that automate the session ritual.
The ritual
Session start: "Load project client-alpha." The AI reads the log and briefs you.
Session end: "Wrap up and log it." The AI writes the summary and next steps.
That's the product.
Tools
read_project_log, append_project_log, save_project_file, list_project_files, read_project_file, search_project, get_recent_activity, list_projects. Plus two MCP prompts: start_session, end_session.
CLI (on the server)
node dist/cli.js admin rotate # new admin key, old one revoked
node dist/cli.js token create|list|revoke <label>
node dist/cli.js export <slug> [dir] # markdown bundle
node dist/cli.js export --all [dir]Security model
Tokens are generated (never defaulted), sha256-hashed at rest, revocable instantly, max 5 active. The admin key (also hashed, rotatable) can mint and revoke tokens but cannot read or write project data. Set ADMIN_API_ENABLED=false for CLI-only token management. All identifiers are allowlist-validated. All data lives in SQLite: no request input ever touches a filesystem path. Non-root container. 60 req/min per token. Activity log keeps who-did-what for 90 days, never content.
Bare-metal install (Debian/Ubuntu)
apt install -y build-essential python3 git curl # better-sqlite3 compiles a native module
# install Node 22, then:
git clone https://github.com/cloudgentic/project-sync /opt/project-sync && cd /opt/project-sync
npm ci && npx tsc
useradd -r -m -s /usr/sbin/nologin projectsync # "sync" is a reserved Debian user; do not use itRun as a systemd service with User=projectsync, DB_PATH=/var/lib/project-sync/project-sync.db, ExecStart=/usr/bin/node dist/server.js. The admin key prints once in journalctl -u project-sync on first start.
Dev
npm install
npm test # 19 tests
npm run devMIT. Want this hosted, multi-project, with a web UI? That's CloudGentic Workspace.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
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
One memory, every AI. A shared, user-owned markdown memory your AI clients read and write over MCP.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
An MCP memory server. One memory your agents share — across models, devices and apps.
Persistent, portable memory for AI assistants — your private memory graph, from any MCP client.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceA self-hosted MCP server that provides AI assistants with a shared, persistent SQLite-backed memory for storing and retrieving project context, decisions, and discoveries. It enables cross-session continuity and team-wide knowledge sharing to keep AI coding tools aligned and informed.3MIT
- FlicenseNot gradedqualityCmaintenanceA remote MCP server that gives AI assistants a shared memory, enabling structured handoffs, journaling, and notes for seamless continuation across sessions.-
- AlicenseNot gradedqualityDmaintenanceA self-hosted MCP server enabling multiple AI coding agents to share state, preserve context across sessions, and coordinate with each other.40Apache 2.0
- FlicenseNot gradedqualityBmaintenanceA self-hosted MCP server providing persistent memory for AI tools, allowing them to remember, recall, and manage information across sessions using a SQLite database.-
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/cloudgentic/project-sync'
If you have feedback or need assistance with the MCP directory API, please join our Discord server