Give your AI agent a real local drive — MeshDrive 2.0 + MCP (stdio)

Описание и применение

MeshDrive 2.0 обеспечивает локальное, изолированное хранилище для AI‑агентов, позволяя безопасно читать и писать файлы через MCP, поддерживая мульти‑тома, квоты и веб‑UI, при этом данные остаются на локальных дисках.

Задача

AI agents can read files only if pointed to random folders, which is fragile, hard to isolate, and lacks durable storage with quotas and web UI.

Результат

MeshDrive 2.0 provides a local-first, isolated storage environment for AI agents, enabling durable, multi-volume storage with quotas and a web UI, while keeping data on local disks and supporting multiple AI clients via stdio MCP integration.

Рабочий процесс

• Install MeshDrive .deb package
• Verify installation with meshdrive doctor and systemctl status
• Create storage backend via meshdrive-tui and mount JuiceFS volumes
• Enable MCP add‑on with sudo meshdrive-addons install mcp
• Run MCP smoke test to confirm tools
• Configure AI client (Cursor, Claude Code, or Hermes) with stdio MCP server JSON
• For remote MeshDrive, set up passwordless SSH and configure SSH command in MCP JSON
• Test MCP integration in AI client by calling health_check and list_storage_backends

Доказательства

MeshDrive is a local-first Linux storage layer (JuiceFS + SQLite metadata) with a terminal UI, loopback Filebrowser, and a free MCP server so agents can list, read, and write files only inside an isolated root

Источник

MeshDrive 2.0 — local-first storage for the agentic era. Built by Vistrix Labs.

Источник