Portable memory for every AI client

One memory across every AI.

Transparent, scoped, and yours — Memora follows you from ChatGPT to Claude to your coding agents.

MCP quick-start

Install one local memory server.

uvx memora-mcp

Memora stores local data in ~/.memora/ by default. Add your own keys when you want hosted models, or leave them out for degraded local recall.

Client configuration

claude_desktop_config.json / .mcp.json

// Claude Desktop (claude_desktop_config.json) / Claude Code (.mcp.json)
{
  "mcpServers": {
    "memora": {
      "command": "uvx",
      "args": ["memora-mcp"],
      "env": {
        "GEMINI_API_KEY": "<optional>"
      }
    }
  }
}

Built for trust

Four pillars for memory that can actually follow you.

Cross-app portability

The MCP server gives Claude, Cursor, coding agents, and other MCP clients the same memory layer without rebuilding context.

Claude

Cursor

Agents

Scoped memory

Project memory stays inside its scope, so project-x memories do not leak into project-y.

project-x

project-y

Trust controls

Memories are visible and manageable with edit, export, and delete controls built into the product surface.

Prefers concise API examples with runnable commands.

BYOK privacy

Use your keys, your machine, and local degraded mode when no cloud provider is configured.

OPENAI_API_KEYoptional

GEMINI_API_KEYoptional

GROQ_API_KEYoptional