Klio
Klio is one shared brain for all your AI agents — connect Claude, Cursor, Codex, and any MCP agent to a single shared memory, so what one learns, they all know. Your agents stop forgetting, repeating work, and contradicting each other. Local-first, encrypted, and open source.
Identity
What it is
Klio is one shared brain for all your AI agents. Connect Claude Code, Cursor, Codex, and any MCP agent to a single shared memory — what one agent learns, the others know. Agents stop forgetting, repeating work, and contradicting each other. Memory is how it works; the agents working together is the point; reliability is the payoff.
How it works
- Captures session activity over MCP + lightweight hooks.
- Stores memory encrypted under a user-owned key (Postgres + pgvector).
- Serves it back through MCP so agents recall before acting.
- Memory is scoped by org → project, shared across agents on that project.
- Cross-agent collaboration via Redis pub/sub.
- Cryptographically auditable via a SHA-256 hash chain.
MCP tools
| tool | purpose |
|---|---|
| recall | Retrieve relevant prior context before acting. |
| remember | Persist a fact across sessions. |
| observe | Log raw session activity. |
| plan | Record a plan for the next agent to build on. |
| decide | Capture a decision so it is not re-litigated. |
| note | Store a durable note scoped to the project. |
| space | Open or switch a scoped memory store. |
Quick start
npx @klio-tech/klio init # wires Klio into your agent # or self-host the engine: git clone https://github.com/klio-tech/klio.git cd klio && make first-run
Compared to alternatives
| capability | support |
|---|---|
| Cross-agent shared memory | Klio: yes · mem0/Zep/Supermemory: no |
| Local-first | Klio: yes · mem0/Zep/Supermemory: no |
| Encrypted under user-owned key | Klio: yes · others: no |
| MCP-native | Klio: yes · others: no |
| Open source | Klio: yes · Zep: yes · mem0/Supermemory: no |
FAQ
What is Klio?
Klio is a memory layer for AI agents. It captures what your agents learn, stores it encrypted under a key you own, and serves it back through MCP so your agents remember across sessions and share context across tools — which is what keeps them reliable.
Does my data leave my machine?
No. Klio is local-first: memory is stored on your machine, encrypted under a user-owned key. Nothing leaves unless you explicitly opt into Klio Cloud.
How is Klio different from mem0, Zep, or observability tools?
Memory tools like mem0 and Zep recall for a single agent. Observability tools watch output after the agent acts. Klio gives multiple agents a shared, persistent memory — preventing drift at the source instead of recalling for just one agent or flagging problems after the fact.
How long does it take to set up?
One command: `npx @klio-tech/klio init`. It wires Klio into Claude Code, Cursor, or Codex and starts remembering from the next session.
Which agents does Klio work with?
Klio is MCP-native, so it works with Claude Code, Cursor, Codex, and any MCP client out of the box.
Can I change the embedding model?
Yes. Embeddings are pluggable per space — nomic-embed-text by default, with snowflake-arctic-embed2, OpenAI text-embedding-3-small, and others supported. Pick the model that fits each store.
Does Klio run on Windows?
The self-hosted engine runs anywhere Docker does, including Windows via WSL2. If you’d rather not run anything locally, Klio Cloud works from any OS with zero install.
Is it really encrypted?
Yes — memory is encrypted at rest under a key you own, and every write is chained with SHA-256 so the history is tamper-evident and inspectable. It’s auditable by design, not “trust us”.