Membrane Engine

Membrane Engine is the runtime that executes your integration logic, handles authentication, and manages data flow between your product and external apps.

Fully Programmable & API-First

Membrane Engine is built with an API-first architecture. Everything you can do through the Membrane Console is available programmatically:

RESTful APIs with SDKs across multiple programming languages
Model Context Protocol (MCP) server exposing 235+ endpoints as tools for AI agents
Command-line interface (CLI) for local development and automation

The CLI enables developers to work in local environments, IDEs, or through AI agents. It synchronizes project files with Membrane's workspace while providing agents with necessary tools and context for integration development.

Monitoring & Debugging

Membrane Engine provides full visibility and control over every integration touchpoint:

Agent Sessions: Check the status of your agentic sessions across any of your integrations
Activity Logs: Get alerts when interfaces stop working or have breaking changes
Error Inspection: Drill down to each API call and fix errors with detailed logs
Package Monitoring: Monitor and debug your Membrane Packages with a few clicks
External Tool Integration: Stream errors to external monitoring tools

See Monitoring & Troubleshooting for details.

Self-Hosting

You can self-host Membrane Engine for complete control:

Data Control: Keep complete control over customer data and meet any compliance standard
No Limits: Remove restrictions on instances, locations, and usage
Full Platform Parity: Engine operates with complete functionality on-premises, identical to cloud version
Easy Deployment: Docker containers and Helm charts provided with production-ready templates for AWS, GCP, and Azure
Minimal Dependencies: Auth0, Mongo-compatible database, S3-compatible storage, and Redis

See Self-hosting for deployment guides.