rajibbora1965/whatsappcoding — explained in plain English
Analysis updated 2026-05-18
Review and approve AI-generated code changes from a phone while away from a computer.
Trigger a coding agent to make changes in a real project folder by sending a WhatsApp message.
Study this repo's proxy design as a reference for reducing token usage sent to an AI coding tool.
Design a similar always-on cloud sync system to keep OAuth credentials fresh for a local development tool.
| rajibbora1965/whatsappcoding | 0xkinno/neuralvault | 0xmayurrr/ai-contractauditor | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | hard | hard | easy |
| Complexity | 5/5 | 4/5 | 2/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
The README is a system architecture description, not a runnable setup, with no included installation steps or code walkthrough.
WhatsApp Coding is a system design document describing a way to control an AI coding assistant on your laptop from your phone, using WhatsApp as the messaging interface. The idea is that a busy person can send a coding request as a WhatsApp message, have an AI agent on their MacBook actually read files, write code, and run shell commands in a real project folder, then approve or reject the result from their phone before anything gets merged. The README describes four main parts working together. First, a dedicated WhatsApp number, kept separate from personal chats, connects through a messaging bridge that mimics a normal desktop browser so the connection is not blocked. Messages sent there trigger a background process on the local computer that performs the actual coding work. Second, a local proxy program intercepts network traffic between the desktop coding tool and its AI provider, stripping out large repeated blocks of text the tool normally sends, which reduces how much text is processed per request, and rebuilding response data so the desktop tool keeps working correctly. Third, that proxy routes expensive, longer AI reasoning requests through a separate flat-rate subscription service instead of a pay-per-use API, to keep costs predictable. Fourth, because a laptop can go to sleep, an always-on cloud virtual machine keeps login credentials refreshed in the background and syncs them back down to the laptop every few minutes. For safety, any code change proposed through this system is shown as a block of text inside the WhatsApp conversation, and the user has to actively confirm they are not driving before a change can be approved or merged. The README ends with a disclaimer noting the various company and product names mentioned are trademarks of their respective owners and that this is an independent project, not officially connected to any of them. This repository is documentation of an architecture rather than a ready-to-run application. It does not include setup instructions, source code walkthroughs, or a license, so it would mainly interest developers curious about the design rather than people looking to install and use it directly.
A system design write-up for controlling an AI coding agent on your laptop from WhatsApp, with a proxy layer and cloud sync to keep it running and affordable.
Mainly TypeScript. The stack also includes TypeScript, Node.js, GCP.
Setup difficulty is rated hard, with roughly 1day+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.