Features
- Provides a managed API for retrieval-augmented generation (RAG)
- Indexes PDFs, CSVs and text for natural language queries
- Cleans and transforms messy data into embeddings automatically
- Dashboard shows retrieved documents, token cost per request and failures
- Scales vector storage and compute automatically
- Swaps between models such as GPT-4o and Claude 3.5 Sonnet with a single toggle
Pricing Details
💵 Pricing (April 2026 Status) Metal follows a "Scale-with-Usage" pricing model, designed to be affordable for startups while offering high-security tiers for large corporations. Plan Price (Approx.) Key Features Free / Developer $0 1 Index; 5,000 documents; Standard API rate limits; Community support. Growth Plan ~$99 / mo 3 Indices; 50,000 documents; SOC2 Compliance; Priority API access. Enterprise Custom Unlimited indices; Millions of documents; Private VPC deployment; SLA. 🎁 Free Version: The Developer Tier is a full-featured "sandbox." It allows you to build a functional prototype with a decent amount of data (5,000 documents) for free, making it ideal for testing a new AI feature or building a personal "Quiet Luxury" knowledge base.