chore: bump version to v1.8.0, update CHANGELOG

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: alvinttang

CHANGELOG.md

@@ -1,5 +1,26 @@
 # Changelog
 
+## v1.8.0 — Cross-Device Memory Sync
+
+### One-Click Sync Setup
+- `sync_enable` MCP tool: auto-detects cloud provider, generates device ID, AES-256-GCM passphrase
+- `sync_pull` MCP tool: pull and apply remote changes from other devices
+- `sync_status` now shows real sync state (device ID, provider, remote devices)
+- CLI: `cortex-mcp-server sync enable` / `sync pull` subcommands
+
+### Improved Install
+- `--ide claude` now uses `claude mcp add` (correct Claude Code registration)
+- New `--ide claude-desktop` for Claude Desktop app
+- Post-install next steps guide
+
+### Documentation
+- Cross-device sync section in README (EN + CN)
+- Japanese (README_JA.md) and Korean (README_KO.md) translations
+- Configurable MCP server name via `CORTEX_SERVER_NAME` env var
+
+### Stats
+- 29 MCP tools, 485+ tests
+
 ## v1.7.0 — Private. Free. Local.
 
 ### Cloud Sync

Cargo.lock

@@ -1,6 +1,6 @@
 [package]
 name = "cortex-core"
-version = "1.7.0"
+version = "1.8.0"
 edition = "2021"
 
 [features]

cortex-core/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cortex-http"
-version = "1.7.0"
+version = "1.8.0"
 edition = "2021"
 description = "HTTP REST API for Cortex memory engine — local-first, no cloud"
 

cortex-http/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cortex-mcp-server"
-version = "1.7.0"
+version = "1.8.0"
 edition = "2021"
 description = "MCP server for Cortex memory engine — gives LLMs persistent memory"
 

cortex-mcp-server/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "cortex-python"
-version = "1.7.0"
+version = "1.8.0"
 edition = "2021"
 
 [lib]

cortex-python/Cargo.toml

@@ -4,7 +4,7 @@ build-backend = "maturin"
 
 [project]
 name = "cortex-ai-memory"
-version = "1.7.0"
+version = "1.8.0"
 description = "Private. Free. Local. — Memory engine for personal AI agents."
 requires-python = ">=3.9"
 license = "MIT"

cortex-python/pyproject.toml

@@ -0,0 +1,121 @@
+# X Thread: How to make your Claude's memory work across all your devices
+
+---
+
+**Tweet 1 (Hook)**
+
+Your Claude forgets everything when you switch devices.
+
+Your MacBook Claude doesn't know what your iMac Claude learned yesterday.
+
+I fixed this. Here's how to give Claude cross-device memory in 2 minutes:
+
+---
+
+**Tweet 2 (Problem)**
+
+The problem:
+- Claude Code stores memory per machine (~/.claude/)
+- Switch laptop? Start from zero
+- Different project on different device? No shared context
+
+Your AI should remember YOU, not just your current machine.
+
+---
+
+**Tweet 3 (Solution intro)**
+
+I built Cortex — a local memory engine that syncs through YOUR cloud storage.
+
+- iCloud, Google Drive, OneDrive, Dropbox
+- AES-256-GCM encrypted (even if your cloud is hacked, memories stay private)
+- 156us writes. 568us search. Pure Rust. 3.8MB.
+
+---
+
+**Tweet 4 (How it works)**
+
+How sync works:
+
+```
+Mac A → writes to iCloud/cortex-sync/
+Mac B → reads from iCloud/cortex-sync/
+
+No server. No API. No account.
+Just your own cloud folder.
+```
+
+Each device writes its own append-only log. No conflicts ever.
+
+---
+
+**Tweet 5 (Setup)**
+
+Setup (2 minutes):
+
+```bash
+# Install
+curl -fsSL https://raw.githubusercontent.com/gambletan/cortex/main/install.sh | bash
+
+# Register as Claude MCP server
+claude mcp add cortex --scope user -- ~/.local/bin/cortex-mcp-server ~/.cortex/memory.db
+```
+
+Then just tell Claude: "Enable cross-device sync"
+
+Claude calls sync_enable, auto-detects your cloud provider, done.
+
+---
+
+**Tweet 6 (What happens next)**
+
+Now when you chat with Claude on Device A:
+
+"I prefer dark mode, use Neovim, and my timezone is Asia/Shanghai"
+
+Switch to Device B, Claude already knows. No setup. No copy-paste. Just works.
+
+Beliefs, facts, preferences, people graph — all synced.
+
+---
+
+**Tweet 7 (Privacy)**
+
+Privacy first:
+
+- Private memories (default) NEVER leave your device
+- Only Shared/Public memories sync
+- End-to-end encrypted with your passphrase
+- Zero telemetry. Zero cloud servers. Verify: grep the source code yourself.
+
+Your memories are yours. Period.
+
+---
+
+**Tweet 8 (Benchmarks)**
+
+vs Mem0 cloud:
+- 528x faster
+- Free (Mem0: $99/mo)
+- 100% local
+- 4-tier memory (Mem0: flat)
+- Bayesian beliefs, people graph, contradiction detection
+
+LoCoMo benchmark (ACL 2024): Cortex 73.7% vs Mem0 66.9% vs OpenAI Memory 52.9%
+
+---
+
+**Tweet 9 (CTA)**
+
+Cortex is MIT licensed, 100% open source.
+
+Give it a try. Give it a star if it helps.
+
+github.com/gambletan/cortex
+
+Now in English, Chinese, Japanese, and Korean.
+
+---
+
+## Hashtags
+#Claude #ClaudeCode #AI #MCP #Rust #OpenSource #AIMemory #Privacy