mcpField
G
Ganglia

Ganglia

Local
github.com/ganglia-toolsDeveloper Tools0Updated 24w ago

About Ganglia

Ganglia is an MCP server that indexes code into a call graph and provides semantic tools for AI coding agents. It supports code navigation, impact analysis, transactional refactoring, and persistent context across sessions.

Key features

  1. 01Call-graph indexing
  2. 02Semantic code navigation
  3. 03Change impact and dependency analysis
  4. 04Transactional refactoring and revert
  5. 05Persistent project context
  6. 06Build, test, and lint tools

Use cases

  1. 01Finding callers and dependencies of a symbol
  2. 02Assessing effects before changing code
  3. 03Applying and reverting cross-file refactors
  4. 04Retaining project context between sessions

MCP support

Ganglia exposes code-graph, analysis, refactoring, build, test, and documentation tools to MCP-compatible clients. The README lists clients including Claude Code, Cursor, Windsurf, Continue, Cline, and Goose.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

No install method listed yet

We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.

Repository

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

Ganglia — code-graph MCP server

npm Site Docs

Ganglia is an MCP server that gives AI coding agents a real understanding of your codebase — not file contents, but structure: who calls what, what breaks if you touch X, where a route ends up in the database.

It indexes your project into a call graph and exposes semantic tools that replace Read / Grep / Glob and slash token usage on real codebases. Works with Claude Code, Cursor, Windsurf, Continue, Cline, Goose — any MCP-compatible client.

What you get

CategoryTools
Readingcode_get, code_signature, code_files, code_toc, code_overview, code_summary, code_structure, code_search, code_grep, code_quick
Navigationcode_callers, code_callees, code_context, code_related, code_trace, code_imports
Analysiscode_impact, code_hotspots, code_dead, code_cycles, code_duplicates, code_breaking_changes, code_diff_context, code_changelog, code_map
Refactor (transactional)code_propagate, code_apply_batch, code_revert, code_history, code_rename_symbol, code_safe_move, code_slice, code_verify_edit
Schema & routingcode_route, code_route_trace, code_schema, code_env, code_type, code_type_usages
Memory & knowledgecode_memory, code_knowledge, code_annotate, code_checkpoint, code_config
Build / test / lintcode_build, code_test, code_lint, code_exec
Deliberationdeliberation_start, deliberation_opinion, deliberation_status, deliberation_result
Smart (LLM-backed)code_smart_read, code_smart_grep, code_smart_diff, code_smart_context
Doc graphdoc_index, doc_list, doc_toc, doc_query, doc_get

Full reference: ganglia.dev/tools · Docs: ganglia.dev/docs

Install

Claude Code (plugin marketplace)
/plugin marketplace add ganglia-tools/marketplace
/plugin install ganglia@ganglia

Includes 5 slash commands and an auto-triggered safe-refactor skill.

Cursor

~/.cursor/mcp.json:

{
  "mcpServers": {
    "ganglia": {
      "command": "npx",
      "args": ["-y", "-p", "@ganglia/cli", "gng", "mcp", "-p", "."]
    }
  }
}
Windsurf

~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "ganglia": {
      "command": "npx",
      "args": ["-y", "-p", "@ganglia/cli", "gng", "mcp", "-p", "."]
    }
  }
}
Continue / Cline / Goose / any MCP client

Same shape — point at the gng mcp -p . command:

npx -y -p @ganglia/cli gng mcp -p .
Direct binary install
npm install -g @ganglia/cli
gng register                 # wires Claude Code automatically

Or grab a platform binary from npm or the download page.

Activation

Free during public beta. One-time setup:

  1. Sign up: https://ganglia.dev/signup (30 seconds, name + email)
  2. Copy your key from https://ganglia.dev/dashboard
  3. gng login <your-key>

Verify: gng whoami shows your tier.

Why use it

Without GangliaWith Ganglia
AI reads whole files to find one functioncode_get("functionName") — 40 lines instead of 800
AI greps blindly across the treecode_grep("pattern", max_results=5) — ranked, structured
AI guesses what breaks if it changes a functioncode_impact(name, depth=2) — severity-scored dependents
Cross-file rename = 47 individual Edit() callscode_propagate → code_apply_batch — one transaction, one revert
Every session starts from scratchcode_memory — persistent context across conversations

Token savings on a real codebase: 80–95% vs raw file-reading patterns.

Languages

TypeScript · JavaScript · Rust · Python · Go · PHP

Repo layout

.claude-plugin/
├── marketplace.json   # Marketplace declaration (one plugin: "ganglia")
├── plugin.json        # Plugin manifest, points at mcp.json
└── mcp.json           # MCP server: npx -y -p @ganglia/cli gng mcp -p .
commands/              # 5 Claude Code slash commands
skills/                # gng-safe-refactor skill
smithery.yaml          # Smithery.ai auto-discovery manifest

Links

The Ganglia source lives in a private repo. This repo is the public-facing distribution — manifests, install instructions, and integration glue. Issues and feature requests welcome here.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.