Stellarion works with any tool that supports the Model Context Protocol (MCP).
The Model Context Protocol is an open standard for connecting AI assistants to external tools and data sources. Stellarion implements an MCP server that exposes its 12 analysis tools.
npm install -g stellarion
Add Stellarion as an MCP server in your client's configuration:
{
"command": "stellarion",
"args": []
}
The exact configuration format depends on your MCP client.
Start your MCP client. It should automatically connect to Stellarion and make its tools available.
When connected, these tools are available:
| Tool | Description |
|---|---|
search_semantic | Natural language code search |
search_pattern | Regex pattern matching |
get_code_element | Extract functions/classes |
analyze_structure | Dependency analysis |
analyze_quality | Code quality metrics |
analyze_impact | Change impact assessment |
index_project | Index for semantic search |
project_info | Project metadata |
generate_report | Quality reports |
find_refactoring | Refactoring opportunities |
extract_endpoints | API endpoint extraction |
batch_analyze | Bulk file analysis |
By default, Stellarion uses the current working directory. You can specify a project path:
{
"command": "stellarion",
"args": ["/path/to/project"]
}
Set environment variables for the server:
{
"command": "stellarion",
"args": [],
"env": {
"OPENAI_API_KEY": "sk-..."
}
}
Once connected, you should be able to:
npm list -g stellarion
stellarion status
Check that: