MCP server that gives AI coding assistants persistent cross-client memory. Facts and conversations stored in SQLite, encrypted with AES-256-GCM, synced to Arweave. No server, no account, recoverable with a 12-word phrase.
Config is the same across clients — only the file and path differ.
{
"mcpServers": {
"sharedcontext": {
"args": [
"-y",
"ai-sharedcontext"
],
"command": "npx"
}
}
}Are you the author?
Add this badge to your README to show your security score and help users find safe servers.
Sovereign, encrypted, portable memory for AI agents.
Run this in your terminal to verify the server starts. Then let us know if it worked — your result helps other developers.
npx -y 'ai-sharedcontext' 2>&1 | head -1 && echo "✓ Server started successfully"
After testing, let us know if it worked:
Five weighted categories — click any category to see the underlying evidence.
No known CVEs.
Checked ai-sharedcontext against OSV.dev.
Click any tool to inspect its schema.
Be the first to review
Have you used this server?
Share your experience — it helps other developers decide.
Sign in to write a review.
Others in ai-ml / developer-tools
Context7 Platform -- Up-to-date code documentation for LLMs and AI code editors
XcodeBuildMCP provides tools for Xcode project management, simulator management, and app utilities.
Copy/paste detector for programming source code, supports 223 formats. AI-ready with token-efficient reporter, skill and MCP server.
A Model Context Protocol (MCP) server and CLI that provides tools for agent use when working on iOS and macOS projects.
MCP Security Weekly
Get CVE alerts and security updates for Sharedcontext and similar servers.
Start a conversation
Ask a question, share a tip, or report an issue.
Sign in to join the discussion.
SharedContext
Sovereign, encrypted, portable memory for AI agents.
SharedContext is a local-first memory layer for LLM agents. It gives Cursor, Claude Code, and Codex persistent memory that you own, synced to Arweave so it follows you across sessions and machines.
Your AI assistant forgets everything the moment a session ends. Project decisions, coding preferences, architectural context — all gone. You re-explain the same things every day.
SharedContext fixes that. It runs as an MCP server that your AI client connects to. Agents store and recall facts through MCP tools. A background watcher syncs your conversation history. Everything is encrypted client-side and uploaded to Arweave. Restore it all on a new machine with a 12-word recovery phrase.
Your AI memory is fragmented:
~/.cursor/projects/~/.claude/projects/None of them talk to each other. None of them persist structured knowledge. None of them let you move to a new machine and pick up where you left off.
SharedContext collapses all of that into one encrypted, portable layer that every client reads from and writes to. One memory. One identity. One place to look.
| Capability | Description |
|---|---|
| Fact memory | Agents store and recall structured facts (preferences, decisions, architecture) across sessions |
| Conversation recall | Retrieve past conversations from any client — "continue the discussion about auth" |
| Encrypted sync | All data is encrypted locally with AES-256-GCM before leaving your machine |
| Arweave persistence | Encrypted blobs are uploaded to Arweave for permanent, censorship-resistant storage |
| Cross-machine restore | 12-word recovery phrase reconstructs your entire context on any machine |
| Conversation sharing | Share a specific conversation via encrypted link — recipient imports with one command |
| Auto-setup | Detects installed AI clients and configures MCP automatically on init |
SharedContext is early (v0.1.x). It works, it's tested, but it's not battle-hardened yet.
Supported clients: Cursor, Claude Desktop, Claude CLI, Codex
Current limitations:
A 12-word recovery phrase deterministically derives a secp256k1 keypair. The phrase + a random salt derive an AES encryption key via Argon2id. The private key is stored encrypted at ~/.sharedcontext/identity.enc. The phrase is stored in the OS keychain for background MCP access.
SQLite is the runtime store. Three tables:
facts — structured memory (key/value with tags, scope, confidence)pending_deletes — tombstones queued for remote syncmeta — version cursors, offsets, sync stateAll reads are local. Fast and offline-capable.
dirty=1upsert / delete)