OpenViSTAI · Agents · Blockchain · Crypto
Menu
MCP & Tooling

Codebase Memory MCP

Codebase Memory MCP is a self-hosted MCP server that parses code repositories into a persistent knowledge graph and provides tools for coding assistants to query code structure, relationships and architecture.

Independent OpenViST project profile, based on linked sources.

Codebase Memory MCP — OpenViST project profile in MCP & Tooling.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when an MCP-compatible coding assistant needs structural, graph-based context from a code repository, or when you want to inspect indexed relationships in the local visualization interface.

Who it is for

  • Developers using MCP-compatible coding assistants
  • Teams building agent workflows for repository navigation, dependency exploration or code review
  • Developers who want a local graph view of indexed code

Documented uses

  • Find and query functions, classes and call chains in a repository.
  • Trace code relationships and explore architecture or potential change impact.
  • Check index coverage, detect dead code and query graph data with Cypher.
  • Inspect indexed code relationships in the built-in graph visualization.
Before you choose

Requirements & limits

The repository is public and MIT-licensed. The README documents prebuilt binaries for macOS, Linux and Windows, with the native install described as requiring no language runtime, Docker or API key.

  • This is an MCP server, so its agent-facing tasks require an MCP-compatible client.
  • The project documents tree-sitter analysis across 162 languages and hybrid LSP type resolution for a more limited set of languages.
  • The graph visualization is optional; the README distinguishes the standard Nix package from the UI-enabled package.
  • The repository reports 17 MCP tools for search, tracing, architecture, impact analysis, index-coverage checks, Cypher queries and other tasks.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the server and connect an MCP client

Install the server with the documented macOS or Linux script, then restart a coding agent so it can use the MCP tools.

Before you start

  • A macOS or Linux environment
  • An MCP-compatible coding agent

Steps

  1. Run the documented install command: curl -fsSL https://raw.githubusercontent.com/DeusData/codebase-memory-mcp/main/install.sh | bash

  2. Restart the coding agent.

  3. Ask the agent to index the project.

Expected outcome

The coding agent can use the server's MCP tools to query indexed codebase information.

Guide limitations

  • The README does not identify client-specific configuration steps for this install path.
  • The install command shown is for macOS and Linux.
Source-based practical guide

Configure an MCP client manually

Add the server executable to the documented Claude Code MCP configuration and check that the tools are visible.

Before you start

  • The codebase-memory-mcp executable installed on the machine
  • Claude Code configuration access

Steps

  1. Add an mcpServers entry named codebase-memory-mcp to ~/.claude.json or the project .mcp.json.

  2. Set command to the executable's path and args to an empty array, as shown in the README example.

  3. Restart the agent.

  4. Use /mcp to check for codebase-memory-mcp and its 17 tools.

Expected outcome

The configured agent lists the server and its MCP tools.

Guide limitations

  • The documented configuration example is specific to Claude Code.
  • The executable path must match its installation location.
Evidence and transparency

Sources & review

Source record details
Source availability
Active
Sources checked
1 October 2026
Original website title
DeusData/codebase-memory-mcp · Official homepage title · retained in its source language