Voyage AI / voyageai.com
High-performance text embedding models optimised for RAG and semantic search applications, offering among the most accurate retrieval embeddings at competitive pricing for AI developers.
Pricing
Free
Free plan
Yes
Category
Developer Tools
Platforms
2
Free plan
Yes
API access
Yes
Open source
No
Platforms
2
Voyage AI is a startup focused exclusively on building best-in-class embedding models for retrieval and semantic search applications. Founded by researchers from Stanford and other leading AI institutions, the company's singular focus produces embedding models with retrieval performance that consistently ranks highly on MTEB (Massive Text Embedding Benchmark) retrieval tasks.
Voyage embeddings include domain-specific models: voyage-finance-2 for financial document retrieval, voyage-code-2 for code search, voyage-law-2 for legal document retrieval, and voyage-3 for general-purpose use. These domain-specialised models outperform general embeddings for their specific domains because they are trained on domain-relevant data.
Context length is a significant embedding model differentiator — Voyage models support up to 32,000 token context length for embedding, enabling long documents like research papers, legal contracts, and financial reports to be embedded as complete documents rather than requiring chunking. This simplifies RAG architectures and can improve retrieval quality.
At $0.006 per million tokens, Voyage is priced significantly below OpenAI Embeddings ($0.02-$0.13/M tokens) and Cohere Embeddings for comparable or better retrieval performance. For high-volume production RAG applications, this pricing differential is significant.
Voyage has secured partnerships with Anthropic (Claude's documentation search uses Voyage), LangChain, and LlamaIndex, validating the model quality among discerning AI application builders.
Voyage AI runs as document analysis llm software built around text and image workflows. Users typically start with a prompt, upload, or connected data source, and the underlying model handles the heavy lifting before returning a result you can refine or export. It's available on web and api, with API access for teams that want to embed it into their own products.
The capabilities that matter most for teams evaluating Voyage AI.
Purpose-trained embedding models for financial documents, code, legal text, and general text producing higher retrieval accuracy for each domain than general-purpose alternatives.
Up to 32,000 token context length enables embedding complete long documents without chunking, simplifying RAG architecture and potentially improving retrieval quality.
$0.006/M token pricing combined with MTEB-competitive retrieval performance produces strong price-performance for high-volume production RAG applications.
Free plan with 50M tokens/month. Pay-as-you-go from $0.006/M tokens. Enterprise custom. Founded by ex-Stanford AI researchers.
Model
Usage-based
Starting price
Free
Free trial
No
OpenAI Embeddings is the most widely adopted — strong ecosystem default. Jina AI (rank 419) provides competitive embeddings with free Reader. Cohere Embed (covered) provides multilingual strengths. Hugging Face provides open source alternatives.
A side-by-side look at the closest alternative in this category.
Key facts about model providers, platforms, and team support.
Model Provider
Voyage AI
Models
voyage-3, voyage-finance-2, voyage-code-2, voyage-law-2
Platforms
Web, API
Deployment
SaaS, API
Integrations
LangChain, LlamaIndex, Anthropic, Weaviate, Pinecone, Qdrant, API
Team Collaboration
No
Launch Year
2023
Compliance signals and data-handling notes as reported by the vendor.
Review Voyage AI's data handling policy. Text and images submitted for embedding processed on Voyage's infrastructure.
Review Voyage AI's privacy policy. Content submitted for embedding is processed on Voyage's servers.
Editorial Verdict
Voyage AI is the best embedding model provider for developers who need high-accuracy domain-specific retrieval (especially finance, code, or legal) or who want the best price-performance ratio for high-volume RAG applications.
Last verified July 24, 2026.
Free plan with 50M tokens/month. Pay-as-you-go from $0.006/M tokens. Enterprise custom. Founded by ex-Stanford AI researchers.
Free plan with 1M tokens. Pay-as-you-go from $0.02/M tokens. Enterprise custom. Reader API free.
Review Voyage AI's data handling policy. Text and images submitted for embedding processed on Voyage's infrastructure.
Review Jina AI's data handling policy. Text and images submitted for embedding processed on Jina's infrastructure. Open source models can be self-hosted.
Review Voyage AI's privacy policy. Content submitted for embedding is processed on Voyage's servers.
Review Jina AI's privacy policy. Content submitted for embedding and reading is processed on Jina's servers for API usage. Self-hosted models keep data private.
Verified reviews from signed-in users, stored in the backend and averaged into this tool's rating.
Sign in to rate Voyage AI and leave a review.
No other reviews yet — be the first to share how this tool performs in practice.
