Engram Rs

Name: Engram Rs
Author: kael-bit

@kael-bit

Engram Rs について

Hierarchical memory for AI agents. Three-layer (buffer/working/core) with automatic decay, promotion, and semantic search.

#memory #ai-agent #mcp #semantic-search #vector-database #rust #sqlite #llm #hierarchical-memory #context-management

基本情報

カテゴリ

メモリとナレッジ

トランスポート

stdio

公開者

kael-bit

投稿者

Kael

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "engram": {
      "command": "npx",
      "args": [
        "-y",
        "engram-rs-mcp"
      ],
      "env": {
        "ENGRAM_URL": "http://localhost:3917",
        "ENGRAM_API_KEY": "",
        "ENGRAM_NAMESPACE": ""
      }
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Engram Rs?

Engram Rs is a Model Context Protocol (MCP) server that gives AI agents persistent, human-like memory. It organizes memories through three cognitive layers — buffer (short-term), working (active knowledge), and core (long-term) — with automatic decay, promotion, and consolidation. Built as a local-first, single-binary Rust server using SQLite, it is designed for developers adding long-term memory to MCP-compatible AI clients.

How to use Engram Rs?

Download the platform-specific Rust binary and run it with any MCP client that supports stdio transport, such as Claude Desktop, Cursor, or Windsurf. The server exposes tools for storing, searching, deduplicating, and managing memories across isolated namespaces, and it can extract session context from LLM conversations.

Key features of Engram Rs

Three-tier memory: buffer, working, and core layers
Semantic search via HNSW vector index
Automatic deduplication and merging of memories
Namespace isolation for multi-agent setups
Session context extraction from LLM conversations
Local-first, single-binary Rust server with SQLite

Use cases of Engram Rs

Give conversational AI agents persistent, long-term memory across sessions
Manage multiple agent identities with isolated memory namespaces
Extract and consolidate key information from ongoing LLM conversations
Maintain context and knowledge without relying on proprietary cloud services

FAQ from Engram Rs

What makes Engram Rs different from other memory systems?

It uses a three-layer cognitive architecture with automatic decay, promotion, and consolidation, mimicking human memory dynamics rather than a flat storage approach.

What runtime or dependencies are required?

None beyond the compiled Rust binary. It is a local-first, single‑binary server that uses SQLite – no external databases, cloud services, or additional runtimes needed.

Where are memories stored?

All memories are stored locally in an SQLite database managed by the Rust server. No data is sent to external services.

What transport and authentication does it use?

It uses the MCP protocol over stdio transport. Authentication is not covered in the README.

What clients does Engram Rs work with?

It works with Claude Desktop, Cursor, Windsurf, and any MCP‑compatible client that supports stdio transport.

「メモリとナレッジ」の他のコンテンツ

Solomd

zhitongblog

A markdown editor — and the bridge to your LLM. Local-first, MIT, ~15 MB. Bundled MCP server lets Claude Code / Codex / Cursor drive your vault directly. 14 AI providers BYOK.

425

Semantic Scholar MCP Server

YUZongmin

A FastMCP server implementation for the Semantic Scholar API, providing comprehensive access to academic paper data, author information, and citation networks.

158