pie connect 3000
# ✓ https://my-app.tunnel.dev → localhost:3000

Why Pipepie?

• End-to-end encrypted — Noise NK protocol. The relay server never sees your data.
• AI-first — auto-detects Replicate, fal.ai, RunPod, Modal, OpenAI webhooks. Pipeline tracing with timeline visualization.
• Self-hosted — your server, your domain, your data. No vendor lock-in.
• Fast — 577 req/s parallel, 2ms overhead. Protobuf + zstd + yamux.
• MCP server built-in — pie mcp gives Claude, Cursor, and AI tools direct access to tunnels, requests, and pipeline traces.
• Beautiful CLI — Dracula theme, interactive forms, 17 commands.
• Zero-config start — pie setup on server, pie login on client, done.

Install

# Homebrew
brew install pipepie/tap/pipepie

# Script
curl -sSL https://raw.githubusercontent.com/pipepie/pipepie/main/install.sh | sh

# Or build from source
git clone https://github.com/pipepie/pipepie && cd pipepie && make build

Quick Start

Server (your VPS, one time)

pie setup

Interactive wizard handles DNS, firewall, TLS (Let's Encrypt auto or Cloudflare), nginx detection, systemd service — everything.

Client (your dev machine)

# Save server connection (key from pie setup)
pie login --server tunnel.mysite.com:9443 --key a7f3bc21...

# Tunnel any local port
pie connect 3000

AI Tool Presets

pie connect --ollama       # Ollama (port 11434, auth enabled)
pie connect --comfyui      # ComfyUI (port 8188, WebSocket)
pie connect --n8n          # n8n workflows (port 5678)
pie connect --tma          # Telegram Mini App (port 5173)

Features

Tunneling

pie connect 3000                    # HTTP tunnel
pie connect 3000 --name my-app     # Stable subdomain
pie connect --tcp 5432              # TCP (databases, gRPC)
pie connect 3000 --auth secret      # Password-protected URL
pie up                              # Multi-tunnel from pipepie.yaml

Webhook Inspection

pie logs my-app --follow --body     # Stream with request/response bodies
pie inspect <request-id>            # Full headers, body, metadata
pie replay <request-id>             # Re-send a captured webhook
pie dashboard                       # Open web UI in browser

AI Pipeline Tracing

Pipepie auto-detects webhooks from AI providers — no configuration needed:

Webhooks from the same pipeline are auto-correlated into traces:

pie traces my-app                   # Terminal timeline view
pie dashboard                       # Web UI with Jaeger-style bars

Or use headers for manual correlation:

curl -X POST https://my-app.tunnel.dev/webhook \
  -H "X-Pipepie-Pipeline: image-gen" \
  -H "X-Pipepie-Step: generate" \
  -H "X-Pipepie-Trace-ID: trace-001"

Multi-Service Config

# pipepie.yaml
server: tunnel.mysite.com:9443
key: a7f3bc21...

tunnels:
  api:
    subdomain: my-api
    forward: http://localhost:3000
  frontend:
    subdomain: my-app
    port: 5173

pipeline:
  name: image-gen
  steps:
    - name: replicate-sdxl
      webhook: /replicate
      forward: localhost:3000/on-image
    - name: fal-upscale
      webhook: /fal
      forward: localhost:3000/on-upscale

pie up

Multi-Account

pie login --server work.example.com:9443 --key abc...
pie login --server personal.example.com:9443 --key def...

pie account                         # List all, see active
pie account use work.example.com    # Switch
pie logout personal.example.com     # Remove

Server Management

pie setup                           # Interactive setup wizard
pie server --config pipepie.yaml    # Start server
pie doctor --config pipepie.yaml    # Diagnose configuration
pie status                          # Tunnel overview
pie status --json                   # JSON output for scripts

Architecture

Client (pie connect)                    Server (pie server)
     │                                       │
     │◄──── Noise NK handshake ────►│
     │       (ChaChaPoly + BLAKE2b)          │
     │                                       │
     │◄──── yamux multiplexing ────►│
     │       (parallel streams)              │
     │                                       │
     │◄──── Protobuf + zstd ──────►│
     │       (binary, compressed)            │
     │                                       │
  localhost:3000              https://sub.domain.com

Noise NK — server authenticated by public key. Know the key = have access. yamux — multiplexed streams, no head-of-line blocking. Protobuf — binary wire format, ~10x smaller than JSON. zstd — bodies >1KB auto-compressed. SSE/streaming — pass-through without buffering (Vercel AI SDK, Ollama compatible).

MCP Server (Claude, Cursor, AI tools)

Pipepie includes a built-in Model Context Protocol server. Your AI tools can inspect webhooks, replay requests, manage tunnels, and debug pipelines — directly from the chat.

# Claude Code (one command)
claude mcp add --transport stdio pipepie -- pie mcp

{
  "mcpServers": {
    "pipepie": {
      "command": "pie",
      "args": ["mcp"]
    }
  }
}

13 tools available:

CLI Reference

Performance

What makes Pipepie different

• Self-hosted & open source — your infrastructure, your data, no third-party traffic inspection
• End-to-end encrypted — Noise NK protocol, the relay server never sees plaintext
• AI-native — auto-detects 6 providers, pipeline tracing, MCP support
• Complete toolkit — tunneling + inspection + replay + dashboard in one binary
• Free forever — no bandwidth limits, no session timeouts, no paid tiers for core features

License

AGPL-3.0 — free to use, modify, and self-host. If you modify and offer as a service, you must open-source your changes.