Two Distinct Claude Surfaces
XMemo offers host-native and remote connection modes for Claude runtimes.
XMemo Claude Plugin
Target Audience: Claude Code and Cowork users
Runtime Contract: claude-plugin · 8 Skills · 7 Hook events · 16 tools · 0 resources
Marketplace Status: 1.0.0 public review candidate
Combines eight professional Skills, a seven-event fail-open checkpoint Hook, and a focused XMemo MCP profile for planning, review, progress recovery, and cross-agent handoffs.
XMemo Claude Connector
Target Audience: Claude users connecting a remote MCP server
Runtime Contract: claude-public · 19 tools · 0 resources
Marketplace Status: Submitted / In review
Exposes account-authorized XMemo memory tools through OAuth 2.0 for remote Claude MCP client integrations.
Representative Workflows
Focused workflows turn live project evidence into durable, retrieval-ready outcomes.
- Brainstorm and plan from evidence: Recall constraints, generate distinct options, challenge assumptions, and turn an approved direction into evidence-gated project and execution plans. Prompt: “Brainstorm the strongest approaches, then turn the approved direction into a project plan.”
- Review plans and audit progress: Compare a proposal or progress report against saved decisions and current project evidence, then separate verified work from drift, gaps, risks, and unsupported claims. Prompt: “Review this plan and progress against what we already decided and what the repository proves.”
- Distill the session and checkpoint milestones: Preserve decisions, rationale, TODOs, verified results, and one exact next action without archiving the raw conversation. The seven-event Hook detects material boundaries and requests a semantic checkpoint without blocking Claude. Prompt: “Distill the durable outcomes from this session and save the verified project checkpoint.”
- Resume from the latest verified state: Reconcile the latest XMemo checkpoint with the live project, surface stale or conflicting assumptions, and continue from one evidence-backed next action. Prompt: “Bring me up to speed from the latest verified checkpoint and the current repository state.”
- Hand work across agents: Preserve provenance, completed work, blockers, open decisions, and the next action so another Claude session or compatible AI agent can continue safely. Prompt: “Prepare an evidence-based handoff for the next agent.”
Choose your Claude setup
Choose either the Claude Code plugin or the remote Claude connector; each has its own setup path.
Claude Code plugin
Version 1.0.0 · public review candidategit clone https://github.com/yonro/xmemo-claude-plugin.gitFor local review: claude --plugin-dir ./xmemo-claude-pluginRun /mcp, select XMemo, and complete browser OAuth
Claude connector
Submitted for directory reviewAdd https://xmemo.dev/mcp as a custom connector in ClaudeComplete browser OAuth for memory:read and memory:writeReturn to Claude and enable XMemo for the conversation
Permissions & Data Boundaries
Clear authorization and storage boundaries for the signed-in XMemo account.
XMemo for Claude requests memory:read and memory:write through browser OAuth. Each call is authorized for the signed-in XMemo account; storage and encryption behavior follows that account's selected security mode.
Developer Resources & Marketplace Status
Official repositories, privacy policy, and technical documentation.
