Exocortex

Name: Exocortex
Author: fuwasegu

@fuwasegu

リポジトリ

Exocortex について

xocortex — an MCP server that gives your AI persistent memory across ALL your projects.

#mcp #ai-memory #knowledge-graph

基本情報

カテゴリ

その他

トランスポート

stdio

公開者

fuwasegu

投稿者

Fuwasegu

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": {
    "exocortex": {
      "command": "uvx",
      "args": [
        "--from",
        "git+https://github.com/fuwasegu/exocortex",
        "exocortex",
        "--mode",
        "proxy",
        "--ensure-server"
      ]
    }
  }
}

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 Exocortex?

Exocortex is a local MCP (Model Context Protocol) server that acts as a developer’s “second brain,” persisting development insights, technical decisions, and troubleshooting records in a centralized knowledge store. It allows AI assistants like Cursor to retrieve contextually relevant memories across all projects, enabling cross-project learning and pattern recognition.

How to use Exocortex?

Install by cloning the repository and running uv sync, then start the server with uv run exocortex. Configure Cursor by adding the appropriate entry to ~/.cursor/mcp.json; options include direct from GitHub (recommended), local installation, proxy mode for multiple Cursor instances, or manual SSE server management on a custom port.

Key features of Exocortex

Fully local: all data and AI processing stays on your machine
Semantic search: find memories by meaning, not just keywords
Knowledge graph: maintains relationships between projects, tags, and memories
Memory links: connect related memories to build a traversable knowledge network
Lightweight and fast: uses embedded KùzuDB and fastembed models
Memory dynamics: smart recall based on recency and frequency
Frustration indexing: prioritize “painful memories” (debugging nightmares)
Web dashboard: cyberpunk-style UI for browsing and health monitoring

Use cases of Exocortex

Cross-project knowledge sharing: lessons from one project immediately available in others
Architecture archaeology: trace the evolution of technical decisions over time via temporal reasoning
Root cause analysis: understand what led to a bug or change by following memory lineage
Pattern recognition: discover common problems and solutions across your entire development history
Automatic memory consolidation: deduplication, auto-linking, and health checks after storing memories

FAQ from Exocortex

Where is data stored?

Your data is stored in ~/.exocortex/ and is preserved regardless of which installation or startup method you choose.

How does proxy mode help with multiple Cursor windows?

Proxy mode automatically starts a single SSE server in the background, allowing multiple Cursor instances to connect to the same server and avoid KùzuDB lock conflicts from concurrent writes.

What is frustration indexing?

Frustration indexing prioritizes “painful memories”—debugging nightmares get boosted in search results, making frequently accessed or emotionally charged insights surface higher.

Are next_actions guaranteed to be executed?

No, execution of next_actions is at the AI agent’s discretion. While the server strongly instructs consolidation, it is not 100% guaranteed due to inherent limitations of the MCP protocol