Skip to main content
Inception’s Mercury models use diffusion to generate tokens in parallel, delivering high-quality responses with low latency. They’re built for responsive AI applications and agent workflows where speed and quality matter. The API is OpenAI-compatible, so you can be up and running in minutes with the tools you already use.

Quick Start

Create an account, grab an API key, and send your first request.

Models & Pricing

Compare Mercury 3, Mercury 2.5, Mercury Voice, Mercury 2, Mercury Edit 2, and Mercury Decide — endpoints, context windows, and pricing.

Capabilities

Streaming, tool use, structured outputs, autocomplete (FIM), and next edit.

API Reference

Explore the REST endpoints for chat, FIM, edit completions, decisions, and models.

Mercury Decide

Classify, route, score, and extract exact spans of text with typed decisions and probabilities.

Integrations

Use Mercury in Cursor, Zed, VS Code extensions, LangChain, Vapi, and more.

Cookbooks

End-to-end guides for search, customer support, voice, SQL, and coding agents.