DEV.co
MCP Servers · DeusData

codebase-memory-mcp

A high-performance code intelligence engine that indexes entire codebases into a persistent knowledge graph in milliseconds, enabling sub-millisecond structural queries for AI coding agents. Shipped as a single static binary with 158 language support and zero runtime dependencies.

Source: GitHub — github.com/DeusData/codebase-memory-mcp
28.2k
GitHub stars
2.1k
Forks
C
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
RepositoryDeusData/codebase-memory-mcp
OwnerDeusData
Primary languageC
LicenseMIT — OSI-approved
Stars28.2k
Forks2.1k
Open issues209
Latest releasev0.8.1 (2026-06-12)
Last updated2026-07-07
Sourcehttps://github.com/DeusData/codebase-memory-mcp

What codebase-memory-mcp is

Tree-sitter AST-based codebase indexer with Hybrid LSP semantic resolution for 9 languages, producing a knowledge graph stored in SQLite. Achieves 120× token reduction and sub-ms query latency through LZ4-compressed in-memory pipeline with Aho-Corasick pattern matching. Exposes 14 MCP tools including architecture analysis, impact mapping, and Cypher queries.

Quickstart

Get the codebase-memory-mcp source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/DeusData/codebase-memory-mcp.gitcd codebase-memory-mcp# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

AI-Assisted Code Refactoring & Navigation

Agents can query codebase structure in sub-ms latency, enabling fast architectural understanding, dead code detection, and cross-service impact analysis without token overhead of file-by-file exploration.

Multi-Service Architecture Analysis

Built-in HTTP route linking, service boundary detection, and infrastructure-as-code indexing (Dockerfile, K8s, Kustomize) make it suitable for distributed systems analysis and dependency mapping.

Rapid Onboarding & Knowledge Graph Exploration

Interactive 3D graph visualization and architecture overview (languages, entry points, hotspots, clusters) accelerate developer onboarding without manual documentation burden.

Implementation considerations

  • Verify agent compatibility: tool auto-detects Claude Code, Cursor, Gemini CLI, Zed, VS Code, Aider, and 5 others; confirm your target IDE is listed before deployment.
  • Plan storage for SQLite knowledge graph; documentation does not quantify disk usage per codebase size—test with representative repo before full rollout.
  • Hybrid LSP semantic resolution requires language-specific servers (Python, TypeScript, Go, etc.); availability and configuration vary by language and must be verified pre-deployment.
  • Graph visualization UI requires Node.js/web stack if UI variant used; standard binary variant has zero dependencies but lacks interactive exploration.
  • Git watcher integration (`auto_watch`) runs background process; performance impact on large monorepos not quantified in documentation.

When to avoid it — and what to weigh

  • Real-Time Collaborative Code Analysis — Designed for local indexing and persistence; no built-in support for concurrent multi-user sessions or real-time collaborative updates described in available documentation.
  • Requirement for Proprietary Language Support Beyond 158 — Tree-sitter coverage is limited to vendored grammars; custom domain-specific languages not in the 158 supported set would require grammar development and recompilation.
  • Strict Air-Gapped Environments Without Binary Distribution — Deployment requires downloading pre-built binaries; compiling from source in pure-C may require toolchain setup not typical in restricted environments.
  • Fine-Grained Real-Time Semantic Analysis Without Indexing Delay — Linux kernel indexing takes 3 minutes; while fast, large codebases incur upfront indexing cost unsuitable for ultra-low-latency interactive scenarios.

License & commercial use

MIT License. Permissive OSI-approved license allowing unrestricted commercial use, modification, and distribution with no warranty. Requires preservation of copyright and license notices in distributions.

MIT is a permissive license explicitly allowing commercial use. No restrictions on proprietary deployment, resale, or integration into commercial products. Attribution (license notice) required but not legal endorsement.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationStrong
License clarityClear
Deployment complexityLow
DEV.co fitStrong
Assessment confidenceHigh
Security considerations

Reads local codebase and writes to agent configuration files—threat model limited to local machine. Binaries signed, checksummed, and scanned by 70+ antivirus engines per release. All processing local; code does not leave the machine. No network calls described. SECURITY.md references present. No hardened runtime isolation, sandboxing, or access controls documented; suitable for trusted development environments.

Alternatives to consider

Sourcegraph

Enterprise code search and intelligence platform with server-based indexing, multi-user support, and web UI; but requires infrastructure deployment and lacks MCP agent integration.

Tabnine / GitHub Copilot (native)

Integrated AI code completion in IDEs with built-in contextual awareness; lightweight but less granular codebase structure querying and no graph-based architecture analysis.

Ctags / Universal Ctags + custom tools

Lightweight tag-based indexing; zero dependencies but lacks semantic resolution, AST parsing across 158 languages, and MCP protocol integration for agent orchestration.

Software development agency

Build on codebase-memory-mcp with DEV.co software developers

Download the latest release, run the one-line install, and start querying your codebase architecture in milliseconds. No configuration required.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

codebase-memory-mcp FAQ

Does it work offline and locally?
Yes. All indexing and querying happens locally; no API keys, no cloud, no network connectivity required. Code never leaves your machine.
What is the indexing time for large repos?
Linux kernel (28M LOC, 75K files) in 3 minutes. Average repos in milliseconds. Actual time depends on file count and language complexity; tree-sitter parsing is the main bottleneck.
Which coding agents are supported?
Auto-detects and configures: Claude Code, Cursor, Codex CLI, Gemini CLI, Zed, OpenCode, Antigravity, Aider, KiloCode, VS Code, OpenClaw, Kiro. Supports any agent implementing the MCP server protocol.
Do I need Hybrid LSP for my language?
No. Tree-sitter AST parsing works for all 158 languages. Hybrid LSP (Python, TypeScript, Go, etc.) adds semantic type information; optional but improves accuracy for those languages if LSP servers are available.

Software developers & web developers for hire

Adopting codebase-memory-mcp is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate mcp servers software in production.

Ready to Index Your Codebase?

Download the latest release, run the one-line install, and start querying your codebase architecture in milliseconds. No configuration required.