codebase-memory-mcp

Name: codebase-memory-mcp
Rating: 88 (1 reviews)
Author: DeusData

MCP Server

DeusData/codebase-memory-mcp

An ultra-fast code intelligence engine that builds persistent knowledge graphs using tree-sitter and hybrid LSP for AI coding agents.

38.8k 3.1kMITUpdated 2026-08-13

Great

View on GitHub Open source file

Overview

Codebase Memory MCP is a high-performance native executable designed to index massive codebases rapidly across 158 languages. It leverages tree-sitter AST analysis and hybrid semantic type resolution to maintain a persistent knowledge graph, dramatically reducing token usage and tool call overhead for LLM coding assistants while providing an interactive 3D visualization UI and a session coordination daemon.

Capabilities

▸Full-repository indexing in milliseconds to minutes
▸Tree-sitter AST parsing for 158 programming languages
▸Hybrid LSP semantic type resolution for major languages
▸Persistent knowledge graph storage and querying (Cypher support)
▸Interactive 3D graph visualization UI at localhost:9749
▸Session coordination daemon for multi-agent synchronization
▸Infrastructure-as-code and K8s manifest indexing

Best for

Exploring large codebases with minimal token consumption, Tracing cross-file function call chains and dependencies, Analyzing architectural impacts and dead code across microservices, Interactively visualizing repository structure via a 3D graph UI

Works with

Claude CodeCursorGitHub CopilotCodex

Evoa Score breakdown= Σ (score × weight)

Task usefulness20%100 → +20.0

実タスクにどれだけ役立つか（機能の豊富さ・用途の明確さ）。 — AIによるcapabilities/use-cases解析

Code quality15%95 → +14.3

実装・指示の品質。 — AIによるSKILL.md/README解析

Maintenance15%100 → +15.0

リポジトリがどれだけ活発に保守されているか。 — GitHub 最終push日時の新しさ

Documentation12%40 → +4.8

ドキュメントの充実度・分かりやすさ。 — README/独自要約の情報量

Security15%80 → +12.0

危険・不審な挙動が無いか。 — AIによるセキュリティレビュー

Originality10%90 → +9.0

ありふれたラッパーではない独自性。 — AIによる独自性判定

Popularity8%100 → +8.0

コミュニティの採用度。 — GitHub Stars/Forks（対数スケール）

Compatibility5%100 → +5.0

対応AIエージェントの広さ。 — AIによる対応エージェント判定

Weighted total88.0 / 100

ライセンス不明/制限あり（Red）のSkillは総合スコアに0.85倍の補正を適用します。ランキングはこのScoreのみで決まり、広告で変わりません。算出方法の詳細 →

Security considerations

Runs 100% locally with no code leaving the machine. Reads local repositories and modifies local agent configuration files during installation. Release archives include checksum verification.