lineai-mcp-server — Claude Code MCP server: dependency impact analysis
An MCP server that feeds Lineai's software dependency data to your AI assistant, so it can tell what a code change will break.
claude mcp add --env LINEAI_SERVER_HOST=<LINEAI_SERVER_HOST> --env LINEAI_USERNAME=<LINEAI_USERNAME> --env LINEAI_PASSWORD=<LINEAI_PASSWORD> --env LINEAI_WORKSPACE_NAME=<LINEAI_WORKSPACE_NAME> --env LINEAI_DEBUG_MODE=<LINEAI_DEBUG_MODE> --transport stdio lineai-mcp-server -- uvx lineai-mcp-serverWhat lineai-mcp-server does
Eight tools come with the server. Two handle impact: lineai-method-impact takes the method you are editing plus an optional class filter for ambiguous matches, and lineai-database-impact maps code against database entities — a column, table or view. The other six talk to Lineai's graph API: node search, dependency traversal from seed_node_ids, shortest-path explanation between two nodes, change-scope validation, owner lookup and a capabilities call.
Every tool accepts optional workspace and materialized_view_id arguments to scope a call. Leave both out and skip LINEAI_WORKSPACE_NAME, and the server falls back to the latest materialized view of its default workspace. MCP schema arguments allow snake_case aliases; request bodies to Lineai use camelCase keys.
Astral UV is required to run it, so uv has to be present before you connect anything. On MacOS uvx can fail to launch inside IDEs such as Cursor, and the workaround is cloning the repo and pointing mcp.json at uv with --directory instead of uvx. Lineai credentials go into LINEAI_SERVER_HOST, LINEAI_USERNAME, LINEAI_PASSWORD and LINEAI_WORKSPACE_NAME; the server cannot authenticate without them.
What it helps with
- find which call sites break after editing a method
- trace how a code change touches database columns and tables
- locate the owner of a node in the dependency graph
- explain the path between two nodes before a refactor
- assess risk for a proposed change from seed nodes
Example prompts for the agent
- Assess the impact of changing the
calculateTotalmethod inOrderService - Which tables and columns does this change touch?
- Explain the shortest path between
from_node_idandto_node_idin the dependency graph
How to install lineai-mcp-server
Platform CLI
claude mcp add --env LINEAI_SERVER_HOST=<LINEAI_SERVER_HOST> --env LINEAI_USERNAME=<LINEAI_USERNAME> --env LINEAI_PASSWORD=<LINEAI_PASSWORD> --env LINEAI_WORKSPACE_NAME=<LINEAI_WORKSPACE_NAME> --env LINEAI_DEBUG_MODE=<LINEAI_DEBUG_MODE> --transport stdio lineai-mcp-server -- uvx lineai-mcp-serverReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from lineai-mcp-server <command>
Platform CLI
codex mcp add lineai-mcp-server --env LINEAI_SERVER_HOST=<LINEAI_SERVER_HOST> --env LINEAI_USERNAME=<LINEAI_USERNAME> --env LINEAI_PASSWORD=<LINEAI_PASSWORD> --env LINEAI_WORKSPACE_NAME=<LINEAI_WORKSPACE_NAME> --env LINEAI_DEBUG_MODE=<LINEAI_DEBUG_MODE> -- uvx lineai-mcp-serverReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from lineai-mcp-server <command>
Config file
{
"mcpServers": {
"lineai-mcp-server": {
"command": "uvx",
"args": [
"lineai-mcp-server"
],
"env": {
"LINEAI_SERVER_HOST": "<LINEAI_SERVER_HOST>",
"LINEAI_USERNAME": "<LINEAI_USERNAME>",
"LINEAI_PASSWORD": "<LINEAI_PASSWORD>",
"LINEAI_WORKSPACE_NAME": "<LINEAI_WORKSPACE_NAME>",
"LINEAI_DEBUG_MODE": "<LINEAI_DEBUG_MODE>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from lineai-mcp-server <command>
Config file
{
"servers": {
"lineai-mcp-server": {
"type": "stdio",
"command": "uvx",
"args": [
"lineai-mcp-server"
],
"env": {
"LINEAI_SERVER_HOST": "<LINEAI_SERVER_HOST>",
"LINEAI_USERNAME": "<LINEAI_USERNAME>",
"LINEAI_PASSWORD": "<LINEAI_PASSWORD>",
"LINEAI_WORKSPACE_NAME": "<LINEAI_WORKSPACE_NAME>",
"LINEAI_DEBUG_MODE": "<LINEAI_DEBUG_MODE>"
}
}
}
}Replace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from lineai-mcp-server <command>
Platform CLI
gemini mcp add -e LINEAI_SERVER_HOST=<LINEAI_SERVER_HOST> -e LINEAI_USERNAME=<LINEAI_USERNAME> -e LINEAI_PASSWORD=<LINEAI_PASSWORD> -e LINEAI_WORKSPACE_NAME=<LINEAI_WORKSPACE_NAME> -e LINEAI_DEBUG_MODE=<LINEAI_DEBUG_MODE> lineai-mcp-server uvx lineai-mcp-serverReplace <…> with your own values. uvx runs the command named after the package; if the README names another one, use uvx --from lineai-mcp-server <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about lineai-mcp-server
How do I install lineai-mcp-server in Claude Code?
Run `claude mcp add lineai-mcp-server -e LINEAI_SERVER_HOST=<LINEAI_SERVER_HOST> -e LINEAI_USERNAME=<LINEAI_USERNAME> -e LINEAI_PASSWORD=<LINEAI_PASSWORD> -e LINEAI_WORKSPACE_NAME=<LINEAI_WORKSPACE_NAME> -- uvx lineai-mcp-server`. Install Astral UV first.
What does lineai-mcp-server do?
It connects Lineai dependency data to your AI assistant. Two tools assess the impact of code changes on source and database, six work with the graph: search, traversal, paths and node owners.
What do I need to run it?
Astral UV for launching the server, plus Lineai credentials: server host, username, password and workspace name.
