StackServicesVoyage AI

Voyage AI

Last updated: Nov 3, 2026

Rationale

Voyage AI is the embedding model provider of Fluid Attacks.

Fluid Attacks chose Voyage AI over other alternatives for these reasons:

  • It is very accurate.
  • It has low latency.
  • It has a model only for code (voyage-code-3).
  • Its API endpoints are easy to use.
  • Its pricing model is based on use.
Model comparison

Alternatives

Fluid Attacks evaluated this alternative and rejected it for this reason.

OpenAI Vector Embeddings

It does not have embeddings only for code.

Usage

Fluid Attacks uses Voyage AI for code embeddings in two components:

  • Matches embeds the titles and descriptions of weaknesses with voyage-3. Then, it runs a cosine nearest-neighbor search on the vectors to match vulnerabilities.
  • Sifts embeds function snippets with voyage-code-3 and sends the vectors to a local candidate classifier. The retrieval of Sifts is agentic and does not use vector search.

Vanta

Learn about Vanta, the official compliance automation tool at Fluid Attacks.

VPC

Learn about VPC, an AWS service employed by Fluid Attacks to host a private network in the cloud.

On this page

RationaleAlternativesOpenAI Vector EmbeddingsUsage