How to Get Your Voyage AI API Key (Best Embeddings for RAG)
Access voyage-3, voyage-code-3, and the top-ranked embedding models on MTEB

If you're building a RAG pipeline with Claude, Voyage AI is the embedding provider Anthropic explicitly recommends. Founded by ex-Anthropic researchers, Voyage specializes in embeddings only - no chat models, no distractions - and consistently tops the MTEB leaderboard. This guide walks you through getting your API key.

Prerequisites

• A free account on dash.voyageai.com
• No credit card required to start (the free tier is genuinely generous)

Step-by-step: creating your API key

1. Go to dash.voyageai.com → sign up or log in
2. Click on the "API Keys" tab in the dashboard
3. Click "Create new key" → give it a descriptive name
4. Copy the key immediately - it starts with pa- and is only shown once
5. Store it in your .env file

What this key unlocks

Voyage AI is focused exclusively on embeddings. With your key, you get access to:

• voyage-3 - the general-purpose flagship, strong across most retrieval tasks
• voyage-3-lite - a faster, cheaper variant for high-volume indexing where top accuracy isn't required
• voyage-3-large - the highest accuracy model, best for precision-critical retrieval
• voyage-code-3 - purpose-built for code retrieval, outperforms general models on technical content
• voyage-finance-2 - specialized for financial documents, earnings reports, regulatory filings
• voyage-multilingual-2 - cross-language retrieval, strong on non-English content

The key insight: domain-specific models (code, finance) outperform general models on their target content by a meaningful margin. If your RAG pipeline has a defined content type, use the specialized model.

Voyage does not offer chat completion or image generation - it's a pure embedding API.

Free tier and pricing

Voyage AI offers 50 million free tokens on signup - this is one of the most generous free tiers in the AI API space. For context, indexing a 1,000-page document corpus typically uses under 5 million tokens. You can build and test a full production RAG pipeline without spending anything.

After the free tier, pricing is per-million tokens and varies by model, with voyage-3-lite being the most cost-effective for high-volume use.

Security - what you need to do

• Never commit pa- keys to git. Use .env, add to .gitignore
• Create separate keys per project - makes revocation easier without breaking other integrations
• Monitor token usage in the dashboard - large document ingestion jobs can be expensive if misconfigured
• Rotate every 90 days as standard hygiene

Using this key with HiWay2LLM

You now have access to the best embedding models available for RAG. Instead of managing the Voyage integration separately in each project, bring your key to HiWay2LLM. You get unified access to Voyage embeddings alongside 200+ LLMs, per-request cost tracking, and budget controls - through a single endpoint. One integration for your entire AI stack.