Connect Your AI to SAGE

Give your AI a permanent memory in under 2 minutes. Works with any AI that supports MCP — Claude, ChatGPT, DeepSeek, Gemini, Mistral, Cursor, local models, and more.

MCP = Model Context Protocol — the standard way AI apps connect to local tools.

SAGE Running

App open or server started

Paste Config

One JSON snippet

Wake Up

Your AI remembers everything

Before you start: Make sure SAGE is running. If you installed the app, just open it. If you built from source, run sage-gui serve. You should see the CEREBRUM at http://localhost:8080/ui/

Open MCP settings

In Claude Desktop: Settings → Developer → Edit Config.

This opens a JSON file called claude_desktop_config.json.

Paste the SAGE config

Add this to the file (replace the entire contents if it's empty):

{
  "mcpServers": {
    "sage": {
      "command": "/Applications/SAGE.app/Contents/MacOS/sage-gui",
      "args": ["mcp"],
      "env": {
        "SAGE_HOME": "~/.sage",
        "SAGE_PROVIDER": "claude-desktop"
      }
    }
  }
}

Windows? Change the command to:
"command": "C:\\Program Files\\SAGE\\sage-gui.exe"

Restart Claude Desktop

Quit and reopen Claude Desktop. You should see a tools icon (hammer) in the chat — click it to verify SAGE tools are listed.

Wake up your AI

Start a new conversation and type:

Take the red pill. Initialize your SAGE memory.

That's it! Claude will initialize its memory and start learning from every conversation. It will remember your projects, preferences, and insights across sessions.

Find your sage-gui path

If you installed from the DMG, it's inside the app bundle. If you built from source, it's wherever you put it.

# macOS (DMG install):
/Applications/SAGE.app/Contents/MacOS/sage-gui

# Built from source:
which sage-gui

Add MCP config to your project

Create a .mcp.json file in your project root:

{
  "mcpServers": {
    "sage": {
      "command": "/Applications/SAGE.app/Contents/MacOS/sage-gui",
      "args": ["mcp"],
      "env": {
        "SAGE_HOME": "~/.sage",
        "SAGE_PROVIDER": "claude-code"
      }
    }
  }
}

Tip: Put this in your home directory for global config, or in each project root for per-project setup. Claude Code picks it up automatically.

Start a new session

Close and reopen Claude Code (or start a new terminal session). SAGE tools will appear automatically.

Wake up your AI

Tell Claude:

Take the red pill. Initialize your SAGE memory.

Done! Claude Code will now build persistent memory across all your coding sessions — debugging insights, project context, your preferences.

Pro tip: Add a CLAUDE.md file to your project with: "Call sage_turn every turn to maintain persistent memory." This ensures Claude never forgets to save its experience.

Option A: ChatGPT Desktop App (Paid plans)

Requires ChatGPT Plus, Pro, Team, or Enterprise. Native MCP support via Developer Mode is available on paid plans only.

Enable Developer Mode

Settings → Beta Features → enable Developer Mode. Then go to Settings → Developer and add an MCP server.

Add SAGE

Paste this MCP config:

{
  "mcpServers": {
    "sage": {
      "command": "/Applications/SAGE.app/Contents/MacOS/sage-gui",
      "args": ["mcp"],
      "env": {
        "SAGE_HOME": "~/.sage",
        "SAGE_PROVIDER": "chatgpt"
      }
    }
  }
}

Windows? Change the command to:
"command": "C:\\Program Files\\SAGE\\sage-gui.exe"

Restart and wake up

Quit and reopen the ChatGPT app. Start a new conversation and say:

Take the red pill. Initialize your SAGE memory. Use sage_turn every reply to remember our conversations.

Important: ChatGPT doesn't have a config file to remind it. Go to Settings → Personalization → Custom Instructions and add: "Always call sage_turn at the start of each reply to maintain persistent memory."

Option B: Chrome Extension (All plans, including Free)

Works on all ChatGPT plans — including free. Uses a Chrome extension to connect SAGE to the ChatGPT website directly in your browser. No desktop app needed.

Install MCP SuperAssistant

Install the MCP SuperAssistant Chrome extension. It bridges MCP tools between your browser and local services.

Add SAGE in the extension

Open the extension sidebar and add a new MCP server:

Command: /Applications/SAGE.app/Contents/MacOS/sage-gui
Arguments: mcp
Environment:
  SAGE_HOME=~/.sage
  SAGE_PROVIDER=chatgpt

Chat on the web

Open chat.openai.com in Chrome and chat normally. The extension handles the rest. Tell ChatGPT:

Take the red pill. You have SAGE memory tools. Use sage_turn every reply to build persistent memory.

SAGE works with any AI that supports MCP (Model Context Protocol). This includes DeepSeek, Gemini, Mistral, Cursor, Windsurf, Cody, LM Studio, and any tool adopting the standard — cloud or local. The setup is the same pattern: point the MCP config at sage-gui mcp.

Find the MCP config for your AI tool

Most MCP-compatible tools have a JSON config file or settings panel for MCP servers. Check your tool's documentation for where to add MCP server configs.

Add SAGE

The config is the same everywhere — just change the SAGE_PROVIDER to match your tool:

{
  "mcpServers": {
    "sage": {
      "command": "/path/to/sage-gui",
      "args": ["mcp"],
      "env": {
        "SAGE_HOME": "~/.sage",
        "SAGE_PROVIDER": "your-tool-name"
      }
    }
  }
}

SAGE_PROVIDER keeps memories scoped per tool. Set it to something descriptive like cursor, windsurf, etc. Facts (verified truths) are shared across all providers automatically.

Restart and initialize

Restart your AI tool, then tell it:

Call sage_red_pill to initialize your memory. Then call sage_turn every reply.

Troubleshooting

"SAGE tools don't appear"

Make sure SAGE is running — check http://localhost:8080/ui/ in your browser. If the CEREBRUM loads, SAGE is running. Also verify the path to sage-gui in your config is correct.

"Connection refused"

SAGE runs on port 8080 by default. Make sure no other app is using that port, and that SAGE was started before your AI tool.

"Initialization fails"

This usually means the embedding provider isn't configured. Open the SAGE app (or run sage-gui setup) to re-run the setup wizard. Pick "Hash" for zero-setup or "Ollama" for semantic search.

Can I use SAGE with multiple AI tools at the same time?

Yes! That's what SAGE_PROVIDER is for. Each tool gets its own memory space. Set different provider names in each config. Facts are shared; personal observations and reflections stay scoped to each tool.

Where is my data stored?

Everything stays on your machine in ~/.sage/. The database is a single SQLite file. Nothing is ever sent to any cloud. If you enabled encryption during setup, all memories are AES-256-GCM encrypted at rest.

Can I switch from ChatGPT to Claude (or vice versa)?

Absolutely. SAGE is model-agnostic — your memories belong to you, not to any provider. Just add the config to your new tool and your AI picks up where the last one left off. Facts are shared across providers automatically.