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
- 01Call-graph indexing
- 02Semantic code navigation
- 03Change impact and dependency analysis
- 04Transactional refactoring and revert
- 05Persistent project context
- 06Build, test, and lint tools
Use cases
- 01Finding callers and dependencies of a symbol
- 02Assessing effects before changing code
- 03Applying and reverting cross-file refactors
- 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.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEGanglia — code-graph MCP server
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
| Category | Tools |
|---|---|
| Reading | code_get, code_signature, code_files, code_toc, code_overview, code_summary, code_structure, code_search, code_grep, code_quick |
| Navigation | code_callers, code_callees, code_context, code_related, code_trace, code_imports |
| Analysis | code_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 & routing | code_route, code_route_trace, code_schema, code_env, code_type, code_type_usages |
| Memory & knowledge | code_memory, code_knowledge, code_annotate, code_checkpoint, code_config |
| Build / test / lint | code_build, code_test, code_lint, code_exec |
| Deliberation | deliberation_start, deliberation_opinion, deliberation_status, deliberation_result |
| Smart (LLM-backed) | code_smart_read, code_smart_grep, code_smart_diff, code_smart_context |
| Doc graph | doc_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:
- Sign up: https://ganglia.dev/signup (30 seconds, name + email)
- Copy your key from https://ganglia.dev/dashboard
gng login <your-key>
Verify: gng whoami shows your tier.
Why use it
| Without Ganglia | With Ganglia |
|---|---|
| AI reads whole files to find one function | code_get("functionName") — 40 lines instead of 800 |
| AI greps blindly across the tree | code_grep("pattern", max_results=5) — ranked, structured |
| AI guesses what breaks if it changes a function | code_impact(name, depth=2) — severity-scored dependents |
Cross-file rename = 47 individual Edit() calls | code_propagate → code_apply_batch — one transaction, one revert |
| Every session starts from scratch | code_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
- Site: https://ganglia.dev
- Docs: https://ganglia.dev/docs
- Tools reference: https://ganglia.dev/tools
- npm: https://www.npmjs.com/package/@ganglia/cli
- AI install recipe: https://ganglia.dev/INSTALL.md
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
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.