Graphify MCP server
Commit Oct 6, 2026Connect your AI assistant to a hosted code knowledge graph to search repositories, trace dependencies, and assess change impact.
Also known as Graphify-Labs/graphify-mcp, graphify-mcp
- Stars
- 3
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 0.1.2
- Last commit
- Oct 6, 2026
- License
- —
What is Graphify MCP?
Graphify connects your AI assistant to a hosted code knowledge graph to search repositories, trace dependencies, and assess change impact.
Summary generated from the project's README
About Graphify MCP server
Connecting an AI assistant to a hosted code knowledge graph allows developers to navigate complex codebases through natural language. By using the Model Context Protocol, this service enables users to search indexed repositories, trace symbol dependencies, and perform static impact analysis to see how code changes might affect a system. It is designed for developers who need to maintain repository context and understand relationships between functions, callers, and linked tests.
The service operates as a hosted MCP server using streamable HTTP transport. Users connect by providing the specific API endpoint to an MCP-capable client, such as VS Code or GitHub Copilot, and authenticate via Graphify OAuth through a web browser. For clients that only support STDIO, a bridge using the mcp-remote package can be used to manage the connection. Access to specific data depends on the user's account permissions, workspace selection, and the completion of repository indexing.
What you can do with it
- 1Discovering repository structuresAsk your AI client to list all available workspaces and repositories within your Graphify account.
- 2Tracing code dependenciesIdentify the specific callers and dependencies associated with a particular code symbol.
- 3Performing impact analysisDetermine which parts of the codebase and which tests might be affected by a specific function change.
- 4Persisting repository knowledgeInstruct the assistant to remember specific repository decisions to maintain context during future interactions.
Written from the project's README. Every use case is checked against a quote from it.
How to install Graphify MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Graphify MCP in Claude Code
$ claude mcp add --transport http graphify-mcp https://api.graphify.com/mcpInstall Graphify MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"graphify-mcp": {
"url": "https://api.graphify.com/mcp"
}
}
}Install Graphify MCP in VS Code
.vscode/mcp.json
{
"servers": {
"graphify-mcp": {
"type": "http",
"url": "https://api.graphify.com/mcp"
}
}
}Install Graphify MCP in VS Code CLI
$ code --add-mcp '{"name":"graphify-mcp","type":"http","url":"https://api.graphify.com/mcp"}'Install Graphify MCP in Codex
$ codex mcp add graphify-mcp --url https://api.graphify.com/mcpInstall Graphify MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.graphify-mcp] url = "https://api.graphify.com/mcp"
Install Graphify MCP in Gemini CLI
$ gemini mcp add -s user --transport http graphify-mcp https://api.graphify.com/mcpInstall Graphify MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"graphify-mcp": {
"url": "https://api.graphify.com/mcp",
"type": "http"
}
}
}Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablegraphify-labs/graphify-mcp is public and reachablegithub.com/graphify-labs/graphify-mcp ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)No licenseFinding
- Commit or release in the last yearLast commit or release 2026-11-13github.com/graphify-labs/graphify-mcp ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeHTTPS endpoint answered 401 (sign-in required)Pass
README
View on GitHubGraphify MCP
Connect your AI assistant to Graphify's hosted code knowledge graph. Search indexed repositories, follow code relationships, assess static change impact, find linked tests, and retain repository knowledge through the Model Context Protocol (MCP).
This repository contains public connection documentation for the Graphify hosted MCP service. The hosted server implementation is maintained privately. This repository does not contain or distribute that implementation.
- Endpoint:
https://api.graphify.com/mcp - Transport: Streamable HTTP
- Authentication: Graphify OAuth through your browser
- Official MCP Registry identity:
io.github.Graphify-Labs/graphify - Documentation: graphify.com/mcp
What you can do
| Workflow | Example request |
|---|---|
| Discover repositories | “List my Graphify workspaces and repositories.” |
| Find implementations | “Find the implementation of this function in my selected repository.” |
| Trace relationships | “Show the callers and dependencies of this symbol.” |
| Assess changes | “What could be affected by changing this function, and which tests are linked to it?” |
| Retain context | “Remember this repository decision, then recall it when we return to this code.” |
Results depend on your account permissions, selected workspace, indexed repositories, and available index data. Static impact analysis and linked-test results help guide investigation; they do not execute tests or prove runtime correctness. Memory operations can persist or modify repository knowledge; review write requests before approving them in your client.
Before connecting
- Sign in or create an account at app.graphify.com.
- Select your workspace and connect a repository you are authorized to use.
- Wait for repository indexing to complete.
- Use an MCP-capable client with support for Streamable HTTP and OAuth. Client/model access and Graphify service plans apply separately.
Connect in VS Code / GitHub Copilot
You can connect directly even if Graphify is not yet visible in the MCP gallery.
- Open the Command Palette and run MCP: Add Server.
- Choose HTTP, enter
https://api.graphify.com/mcp, and name the servergraphify. - Choose the desired configuration scope and start the server.
- Complete the Graphify browser sign-in and authorize the connection.
- Confirm Graphify tools are available. Open an agent chat with MCP tools enabled and ask it to list your Graphify repositories.
Alternatively, use this VS Code configuration in .vscode/mcp.json:
{
"servers": {
"graphify": {
"type": "http",
"url": "https://api.graphify.com/mcp"
}
}
}
No password or access token belongs in this file. Let the client complete OAuth and manage its credentials. See VS Code's MCP instructions for current UI and configuration options.
Connect in other MCP clients
Client-specific packages and setup instructions:
- OpenClaw / ClawHub — published MCP bundle and usage skill; see the local bundle notes.
- Nous Research Hermes Agent — native OAuth configuration and an official-catalog contribution candidate.
These are connection packages for the hosted service. Their inclusion here does not mean every platform has approved or shipped a catalog listing.
For clients with a remote MCP setup form, enter https://api.graphify.com/mcp as the server URL and complete browser OAuth. Configuration formats differ by client.
Clients that accept an mcpServers configuration may use:
{
"mcpServers": {
"graphify": {
"url": "https://api.graphify.com/mcp"
}
}
}
For a STDIO-only client or one without compatible remote OAuth support, a bridge configuration is available:
{
"mcpServers": {
"graphify": {
"command": "npx",
"args": [
"-y",
"mcp-remote@0.14.3",
"https://api.graphify.com/mcp",
"--transport",
"http-only"
]
}
}
}
The bridge requires Node.js and npm on the client's PATH. Enabling it downloads and runs the pinned third-party mcp-remote package, which handles browser OAuth and stores credentials locally. Prefer your client's native remote connection when it supports Graphify's OAuth flow.
Verify the connection
- Confirm the MCP client reports the server as connected and exposes Graphify tools.
- Ask it to list your workspaces and repositories.
- Select a populated, indexed repository and request a known function.
- Ask for that function's callers or linked tests and check the returned evidence.
Tool discovery alone does not verify repository access or the accuracy of a particular query result.
Troubleshooting
- Server missing from a gallery: add the endpoint directly. Official MCP Registry publication and downstream gallery inclusion are separate steps.
- Sign-in fails: check that your client supports Graphify OAuth, retry its sign-in flow, or use the bridge configuration where supported.
- No repositories or empty results: confirm workspace access and indexing completion in Graphify.
- Node.js or npx not found: ensure Node.js and npm are available on the MCP client's PATH when using the bridge.
- Tools connect but chat cannot run them: check model access and tool enablement in your AI client.
Support and policies
Do not post credentials, tokens, private repository contents, or sensitive logs in public issues. Use the support channel for account-specific assistance.
Graphify MCP questions
How do I add Graphify MCP to Claude Code?
Run claude mcp add --transport http graphify-mcp https://api.graphify.com/mcp in your terminal. Add --scope user to make it available in every project.
How do I add Graphify MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What can I do with Graphify's MCP service?
You can discover repositories, find function implementations, trace symbol dependencies, assess the impact of code changes, and retain repository context.
Does Graphify execute my tests?
No, the service provides static impact analysis and identifies linked tests to guide your investigation, but it does not execute them.
What if my MCP client only supports STDIO?
You can use a bridge configuration with the mcp-remote package to connect to the hosted HTTP service.