Skip to main content

Introduction

Air Pipe is a platform for shipping high-performance APIs, workflows, integrations, and AI/MCP services from a prompt, declarative YAML, or a marketplace pack.

Deploy on managed cloud or a single self-hosted binary. Every HTTP and MCP surface can ship with OpenAPI docs, Prometheus metrics, and OpenTelemetry tracing — observability by default, not a black box.

What you can build​

  • REST/CRUD APIs over Postgres, MySQL, SQLite, MongoDB, and more
  • Webhook pipelines (Stripe, GitHub, custom) with durable state and retries
  • MCP tool servers so AI agents call your data safely
  • RAG chatbots, LLM gateways, scheduled jobs, multi-tenant backends

Configs are real deployable services with schema validation — not throwaway LLM sketches.

How to start​

Pick the path that matches how you like to work:

15-minute tracks

Three Peerlist-friendly paths: MCP Postgres, REST API, or Stripe webhooks

Beginner tutorial

Hand-written YAML walkthrough — setup through workflows

Install

Managed cloud, binary, container, Linux, macOS, Windows

Configuration

Interfaces, actions, globals, interpolation, and schema reference

Marketplace

Fork a ready pack — REST, MCP, RAG, webhooks, SaaS starters

Examples

Worked examples across databases, HTTP, AI, and marketplace packs

Why teams pick Air Pipe​

  1. Ship fast — prompt, YAML, or marketplace pack; go live in minutes.
  2. See everything — OpenAPI (docs: true), Prometheus (expose_metrics: true), OpenTelemetry on the binary.
  3. AI-native both ways — Build with AI for config generation; MCP tools to expose your APIs to agents.
  4. Run anywhere — app.airpipe.io managed, or self-host with one binary — no lock-in story.
  5. Real backend primitives — databases, webhooks, durable state, scheduling, access control, flow control.

Not sure where to begin?​

Start with the 15-minute onboarding tracks. Each track ends in something you can hit with curl, an MCP client, or Stripe CLI — and that you can inspect with OpenAPI and metrics.