pa-embed

# pa-embed

Generate text embeddings for semantic search using Ollama's embedding models. Outputs embedding vector with metadata.

## Installation

This is a CmdForge tool. It should be installed and available in your PATH via CmdForge.

```bash
# Verify installation
which pa-embed
```

## Usage

This tool accepts text via stdin and outputs a JSON object containing the embedding vector and metadata.

```bash
# Basic usage
echo "What is the capital of France?" | pa-embed

# Use a different model
echo "Hello world" | pa-embed --model mxbai-embed-large

# Use a custom Ollama endpoint
echo "Sample text" | pa-embed --endpoint http://192.168.1.100:11434

# Chain with other tools
cat query.txt | pa-embed | jq '.embedding | length'
```

## Arguments

| Flag | Default | Description |
|------|---------|-------------|
| `--model` | `nomic-embed-text` | Ollama embedding model to use |
| `--endpoint` | `http://localhost:11434` | Ollama API endpoint URL |

## How It Works

1. **generate_embedding**: Reads text from stdin, sends it to the Ollama API endpoint with the specified model, and generates an embedding vector
2. Outputs the result as JSON containing the embedding and associated metadata

The tool uses Ollama's embedding API to convert text into dense vector representations suitable for semantic search, similarity comparison, and retrieval-augmented generation (RAG) workflows.

## Dependencies

None. Requires a running Ollama instance at the specified endpoint with the embedding model available.