Getting Started

Up and running in
5 minutes

No Docker. No cloud account. No configuration required to start.

macOS / Linux / Windows Single binary No external deps

Install MuninnDB

One command installs the binary on macOS, Linux, or Windows:

bash

# macOS / Linux
curl -fsSL https://muninndb.com/install.sh | sh

# macOS (Homebrew)
brew install scrypster/tap/muninn

# Windows (PowerShell)
irm https://muninndb.com/install.ps1 | iex

Initialize and connect your AI tools

One command sets everything up — connects Claude Desktop, Cursor, VS Code, or Windsurf, generates a bearer token, and starts all services:

bash

muninn init

# Guided setup wizard — connects Claude Desktop, Claude Code, Cursor, OpenClaw, Windsurf, Codex, VS Code
# [1/3] Which AI tools would you like to connect?
# [2/3] Secure your MCP endpoint with a bearer token? [Y/n]
# [3/3] Start MuninnDB now? [Y/n]
#
# muninn started (pid 12345)
#   MBP  :8474   binary protocol
#   REST :8475   JSON API
#   gRPC :8477   gRPC API
#   MCP  :8750   AI tool integration
#   UI   :8476   http://localhost:8476

Web dashboard

Visit http://localhost:8476 for the visual dashboard — decay charts, relationship graphs, live activation log.

Store your first memory

Using the Go or Python SDK (REST on port 8475):

package main

import (
    "context"
    "fmt"
    "github.com/scrypster/muninndb/sdk/go/muninn"
)

func main() {
    ctx := context.Background()
    client := muninn.NewClient("http://localhost:8475", "your-token")

    // Store a memory engram
    engID, err := client.Write(ctx, "default",
        "user prefers dark mode",
        "Always render UI in dark theme for this user",
        []string{"preference", "ui"})
    fmt.Printf("Stored engram: %s\n", engID)
    _ = err
}

import asyncio
from muninn import MuninnClient

async def main():
    async with MuninnClient("http://localhost:8475", token="your-token") as client:
        # Store a memory engram
        eng_id = await client.write(
            vault="default",
            concept="user prefers dark mode",
            content="Always render UI in dark theme for this user",
            tags=["preference", "ui"],
        )
        print(f"Stored engram: {eng_id}")

asyncio.run(main())

Activate relevant memories

// Find the top 5 most relevant memories for a context
results, err := client.Activate(ctx, "default", []string{"what does the user want?"}, 5)

for _, r := range results.Engrams {
    fmt.Printf("%.2f — %s\n", r.Score, r.Concept)
    fmt.Printf("  Why: %s\n", r.Why)
}

// Output:
// 0.94 — user prefers dark mode
//   Why: BM25 match (0.78) + Hebbian boost (0.16)

# Find the top 5 most relevant memories for a context
results = await client.activate(
    vault="default", context=["what does the user want?"])

for r in results.engrams:
    print(f"{r.score:.2f} — {r.concept}")
    print(f"  Why: {r.why}")

# Output:
# 0.94 — user prefers dark mode
#   Why: BM25 match (0.78) + Hebbian boost (0.16)

That's it

MuninnDB automatically handles decay, Hebbian learning, and association building from here. Your memories improve the more they're used. Use muninn shell to explore interactively.