CLI Reference

Embenx provides a powerful Command Line Interface (CLI) for benchmarking, setup, and ecosystem interaction.

General Commands

info

Display system information and status of installed vector backends.

embenx info

check

Verify environment health and dependency availability.

embenx check

setup

Setup the environment and pull specific models (for Ollama).

embenx setup --model ollama/nomic-embed-text --pull

Benchmarking

benchmark

Run performance benchmarks on local Parquet files.

embenx benchmark data.parquet --indexers faiss,usearch --max-docs 1000

grand-benchmark

Run comprehensive benchmarks across all datasets in the Retrieval Zoo and generate a Grand Technical Report.

embenx grand-benchmark --indexers faiss,scann,hnswlib

Agentic & Visual

mcp-start

Start the Model Context Protocol (MCP) server to connect AI agents.

embenx mcp-start

explorer

Launch the Streamlit web UI for data visualization and HNSW graph inspection.

embenx explorer

Retrieval Zoo

zoo-list

List available pre-indexed collections in the cloud zoo.

embenx zoo-list

zoo-load

Download and load a pre-built collection into local cache.

embenx zoo-load squad-v2

Maintenance

cleanup

Remove temporary benchmark artifacts and database files to free up space.

embenx cleanup

init-skill

Initialize the Embenx skill for Gemini CLI sub-agents.

embenx init-skill