Promote and share
Cloud
Every client you connect reads and writes the same datapack. To see how this works, connect a second client.
Install using your agent
Use with Cursor, Claude Code, Claude Desktop, Codex, Kiro, OpenCode
Now ask the new client "What does Meko remember about my vacation preferences?" It answers from the memory your first agent created.
To bring your team in, open your datapack and click Share, then Invite People. Enter email addresses and choose a role:
- Viewer reads the datapack.
- Contributor also adds memories.
- Maintainer also edits the knowledge base, promotes memories, and manages members.
Click Send invite to send.
To share a link instead, click Share, then Create share link.
Each teammate connects their own agents and picks the datapack you shared. When a maintainer promotes a memory from Learnings in the portal, it becomes collective memory that every member's agents retrieve, so one person's agent learns a standard and everyone's follows it. See Learnings and Manage access.
Local
Agents you connect on this computer share its datapacks. Connect a second one and ask it the same question:
meko agent connect cursor
Propose a 5 day vacation for me.
To invite other people, link your local Meko to Meko Cloud. Agents stay connected to Meko on your computer, which sends each call to the side that has the datapack. New datapacks are created in Meko Cloud.
meko cloud-link
Sign in with the browser, or paste an API key. In the web UI, you can link to Meko Cloud using Settings > Meko Cloud.
Datapacks already on your computer stay there. A datapack you create after linking is in Meko Cloud, so create one to share:
meko-cli create_datapack team-notes
Then open it in the portal and invite people with the Share button.
What you just did
You connected an agent to a datapack, gave it knowledge it could search, watched it learn a preference that a fresh session recalled, and shared that context with a team. The same loop works for any agent that speaks MCP.
Next steps
Context management
Memory, knowledge, conversations, and artifacts in depth.
Connect to an AI agent
Per-client setup, installer options, and uninstall.
Sharing and governance
Roles, share links, and who can read what.
Build agentic apps
Patterns for agent teams that hand off work and learn together.