Anvil JS is a Node.js backend framework designed for AI and GenAI developers, offering Express-level flexibility and Next.js-style file-based routing. It features a native agentic layer with built-in MCP auto-exposure, a tool registry, agent orchestration, observability, and durability. This framework streamlines the development of AI applications by integrating essential components directly into its core, eliminating the need for separate codebases or manual configurations for features like MCP servers, tool schemas, or tracing infrastructure.

• File-based Routing: Maps folder structures to HTTP routes, supporting dynamic parameters, catch-alls, and middleware scoping.
• Native Agentic Layer: Includes MCP auto-exposure, a tool registry, agent orchestration, observability, and durability.
• Multi-Protocol Support: Exposes any route to speak REST, MCP, and A2A protocols with a single handler.
• Compile-time Safety: Validates parameter names, schema compatibility, and MCP serializability during the `anvil build` process.
• Agent-Native Primitives: Provides built-in features for defining agents, LLM clients, guardrails, durable checkpointing, and human-in-the-loop interactions.
• Built-in Observability: Traces agent runs to a local SQLite store and supports OpenTelemetry export.
• Simplified Setup: Allows for the creation of a first Anvil app with file-based routes and an agent endpoint in under 5 minutes.

Anvil JS simplifies the process of building AI products on Node.js by providing a unified framework that handles HTTP routing, agent management, LLM integrations, and observability, reducing the complexity of stitching together multiple tools and libraries.