Function embed_directory

Embeds an entire directory with cost limit protection.

This is the recommended high-level function for embedding a directory. It performs the following steps with safety checks:

1. Chunks the directory structure
2. Calculates statistics and estimated cost
3. Checks cost against limit (aborts if over limit)
4. Generates embeddings using batch API

Throws

Error if estimated cost exceeds cost limit

Example

// Embed directory with default $1 limit
try {
  const result = await embed_directory('/path/to/project');
  console.log(`Success! Generated ${result.embedStats.successfulEmbeddings} embeddings`);
  console.log(`Actual batches used: ${result.embedStats.totalBatches}`);
} catch (error) {
  console.error('Embedding failed:', error.message);
}
Copy

Example

// Custom cost limit and file filtering
const result = await embed_directory('/path/to/src', {
  costLimit: 0.10,  // 10 cents max
  fileExtensions: ['.ts', '.tsx'],
  delimiter: /\n{3,}/g
});
Copy

Example

// Use larger model with custom dimensions
const result = await embed_directory('/path/to/project', {
  costLimit: 5.0,
  model: 'text-embedding-3-large',
  dimensions: 512,
  costPerToken: 0.00000013  // Update for larger model pricing
});
Copy