MemoryPort
Permanent Memory for any LLM
Persistent memory across every LLM. Every session picks up where you left off. Local-first and open source.
Local-first
Runs on your machine. LanceDB vector index, zero cloud dependency. Your data stays yours.
Permanent backup
Optional Arweave storage. Survive disk failures, sync across devices, rebuild from chain.
Encrypted
AES-256-GCM with Argon2id key derivation. Even if you use cloud backup, only you can read it.
Intelligent retrieval
Three-gate system with query expansion. Your AI gets the right context, not everything.
Multi-tool memory
Claude Code, Cursor, Ollama, OpenAI — one memory across all your AI tools.
Open source
MIT licensed. Full Rust codebase. Read every line, self-host, or contribute.
Claude Code
MCP Server
Cursor
MCP Server
Open WebUI
Ollama proxy
Ollama
Native API proxy
ChatGPT
OpenAI proxy
Any MCP client
Standard protocol
Free
Full AI memory on your machine
- Unlimited local storage
- Claude Code integration (MCP)
- Cursor integration
- Ollama & Open WebUI support
- Knowledge graph visualization
- Session analytics
- Encrypted at rest
Pro
Cloud backup + cross-device sync
- Everything in free
- Memories available forever, even after you stop paying
- 1 GB storage included
- Cross-device sync
- Cold start rebuild from chain
- Priority support
Never repeat yourself to an LLM again
Start building with persistent memory. With MemoryPort, your AI remembers every conversation you've ever had.