Ditto V2: a shared memory layer across MCP clients
Built on Bittensor, Ditto V2 links Claude, Cursor, Codex and Grok through MCP with importable history, approval-gated actions and $1 free credit.
Switching between AI tools shouldn't mean starting from scratch every time. Your research is in Claude, your code is in Cursor, and your notes are somewhere else. The context gets scattered, and you're left connecting the dots yourself. @heydittoai (SN118) is tackling this with Ditto V2, a shared memory layer that connects your AI tools and helps them work together. What you can do with Ditto V2: • Shared memory: Connect Claude, Cursor, Codex, Grok, and other MCP clients. • Bring your context: Import history from ChatGPT, Claude, Notion, Obsidian, and local notes. • One connected workspace: Bring messaging apps, GitHub, Google Workspace, documents, and code into one memory system. • Get things done: Run scheduled workflows, build and deploy small apps, and automate website tasks. • Stay in control: Actions that send, spend, or delete require your approval. You can also switch models without losing context and export your memory through a signed Memory Passport. Getting started is free at http://heyditto.ai, with no card required and $1 in free credit for new accounts. Built on Bittensor, Ditto V2 sources inference and storage from other subnets while miners compete to improve its memory layer. If you're tired of managing context across multiple AI tools, this is worth exploring.