LlamaIndex Embeddings Integration: Fireworks

This integration provides support for Fireworks AI embedding models with LlamaIndex.

Fireworks AI offers fast and efficient inference for embedding models. Sign up at fireworks.ai to get an API key.

Installation

bash

pip install llama-index-embeddings-fireworks

Usage

python

from llama_index.embeddings.fireworks import FireworksEmbedding

# Initialize the embedding model
embed_model = FireworksEmbedding(
    api_key="your-api-key",  # or set FIREWORKS_API_KEY env var
    model_name="nomic-ai/nomic-embed-text-v1.5",
)

# Get embedding for a single text
embedding = embed_model.get_text_embedding("Hello, world!")

# Get embeddings for multiple texts
embeddings = embed_model.get_text_embedding_batch(
    ["Hello, world!", "How are you?"]
)

# Get query embedding
query_embedding = embed_model.get_query_embedding("What is machine learning?")

Configuration

Parameter	Description	Default
`api_key`	Fireworks API key (or set `FIREWORKS_API_KEY` env var)	None
`model_name`	Embedding model to use	`nomic-ai/nomic-embed-text-v1.5`
`dimensions`	Output embedding dimensions (if supported by model)	None
`embed_batch_size`	Batch size for embedding requests	10
`timeout`	Request timeout in seconds	60.0
`max_retries`	Maximum number of retries	10

Environment Variables

FIREWORKS_API_KEY: Your Fireworks API key
FIREWORKS_API_BASE: Custom API base URL (optional)