AI coding tools forget everything.
We fix that.

ContextSync gives every AI tool a shared memory — local-first, privacy-first. One setup, 10 tools connected.

Start Free View Pricing
10 AI tools · Local SQLite · MCP open standard

Every AI tool starts from zero

Developers use 3–5 AI coding tools daily. Every one starts blank — no memory of what you taught yesterday.

  • Wasted time — Re-explaining context to each tool, ~4 min every morning
  • Inconsistent responses — Tools lack shared context, answers vary
  • Duplicated rules — CLAUDE.md, .cursorrules, AGENTS.md, GEMINI.md…
  • Token costs — Repeated explanations eat your context window and budgets
  • Locked to one machine — Your AI context doesn't travel with you

How we fix it

🔗

Unified Memory

SQLite + FTS5 search. 4 MCP tools: get_memories, save_memory, get_rules, list_memories. Any MCP tool connects automatically.

📝

Rule Sync

One rules.md → CLAUDE.md, .cursorrules, AGENTS.md, .windsurfrules, GEMINI.md + more. Daemon auto-updates all files.

🔒

Privacy-First

Local by default. AES-GCM encrypted tool list. Cloud sync is opt-in. We never read your content.

Built on MCP

The Model Context Protocol is an open standard by Anthropic for connecting AI tools to external data. ContextSync implements an MCP server that exposes your project memory as tools any MCP-compatible AI can use.

Each AI tool launches ContextSync via stdio. The MCP server connects to a shared local SQLite database with FTS5 full-text search. A background daemon watches rules.md for changes and syncs memories to cloud for Pro users.

Supported tools

Aider Claude Code Cline Codex CLI Continue Cursor Gemini CLI GitHub Copilot Replit AI Roo Code Windsurf

The numbers

~4 min
saved per session
~55%
token cost reduction
10+
tools supported

Roadmap

  • Team Pro — Seats, invites, shared-context workflows (join waitlist)
  • Memory stats — See how often saved memories saved you from re-explaining
  • Rule drift audit — Weekly scan flagging stale rules referencing deleted files
  • Freelancer profiles — Per-client memory profile and billing reports
  • More integrations — Expanding beyond the current 10 tools

Who's behind this?

Yang Qing — independent developer. Built ContextSync after years of frustration with the same memory-loss problems across every AI coding tool. The solution had to be tool-agnostic.

support@yangqing.one blog.yangqing.one ↗