Perplexity Launches Two Embedding Model Families for Web-Scale Retrieval.

Perplexity has announced two new embedding model families: pplx-embed-v1 and pplx-embed-context-v1. According to the company, both APIs are designed for real-world, web-scale retrieval and are positioned as state-of-the-art embedding models.

Key details from the announcement:

• Perplexity released two separate embedding model families rather than a single model.
• The new APIs are named pplx-embed-v1 and pplx-embed-context-v1.
• The company says the models are built specifically for retrieval tasks at web scale.
• The launch appears focused on developers building search, ranking, and retrieval-augmented applications.
• Perplexity shared a product page for the release at pplx.ai/pplx-embed.

For small business owners and builders, this expands the options for powering AI search and knowledge retrieval features inside products, internal tools, and customer support systems. Teams using NoCode or lightweight AI stacks may also see more choices for connecting embedding APIs to vector databases and retrieval workflows.