Inception LabsMercury API Documentation

Lightning-fast AI inference with OpenAI-compatible API. Chat completions, code editing, streaming, tool use, and IDE integrations.

Mercury 2 Chat

128K context, tool calling, structured outputs, streaming, and diffusion visualization. From $0.25/1M input tokens.

Fill-in-the-middle autocomplete, apply edits, and next-edit predictions for code. 32K context window.

Near-real-time responses with reasoning_effort=instant. Perfect for voice assistants, chatbots, and low-latency workflows.

Works with Cursor, Cline, Roo Code, Kilo Code, Zed, OpenCode, and OpenClaw out of the box.

Watch outputs refine in real-time with the unique diffusion visualization mode.

Drop-in replacement for OpenAI API. Use existing client libraries, LangChain, and frameworks.