FixFramer: A SaaS platform for automatic reverse engineering and deployment of React applications

FixFramer: a SaaS platform for automatic reverse engineering and deploying React applications

Complex architecture, circumventing Vendor Lock-in, binary patching, and custom billing. Full-stack solution "turnkey".

Problem (The Challenge)
Task: Framer — a closed ecosystem.
Sites are tightly bound to their infrastructure, which creates risks of blocking (Roskomnadzor) and the inability to host on their own servers.

Difficulty: Export is not just a "download".
Framer is a complex React component graph, spread across a CDN, with binary configs and dynamic imports. The challenge was to "pull" the site out of the ecosystem while maintaining 100% interactivity.

Engineering Solutions (The Core / Hardcore Engineering)

Deep Reverse Engineering: Developed a core that reconstructs the dependency graph of the application. Parsing ES Modules, intercepting dynamic import() and new URL(...) to create an offline version of the site.
Binary Patching (Low-level): Implemented a bin-patching algorithm for .framercms files. Replacing CDN links inside binaries while preserving the original file length (padding) and byte offsets. This allowed redirecting requests to our cloud without breaking the proprietary client parser.
Smart Synchronization (Diff Loading): Implemented a differential loading algorithm on S3. The system downloads and updates only the changed assets, removing stale assets. This saves traffic and speeds up deployment by 10 times.
Network Layer: Working at the urllib level with manual management of SSL contexts and Brotli/Gzip compression to emulate the behavior of a real browser and bypass bot protection.

Infrastructure and SaaS integration (The Business Logic)

Complex Billing: Integration of YooKassa with support for Proration (automatic recalculation when changing plans), automatic payments (recurring payments), slot management, and fraud protection.
Infrastructure:
- Backend: Python (FastAPI/Django) + Redis (protected cache).
- Async Tasks: Celery for managing heavy parsing tasks in the background.
- Database: Supabase (PostgreSQL).
Client Functionality:
- Binding custom domains with automatic DNS record verification.
- Secure authorization (Auth + Reset password).
- Automatic SSL generation and deployment.

Result (The Outcome)

A fully automated SaaS. The user enters a link to a Framer project -> within 2 minutes receives a standalone website deployed on independent cloud in the RF.

Stack: Python, Java-script, Redis, Celery, S3, Docker, Supabase, React, Type-script.