remote-dev-ai

Description

Remote‑Dev‑AI delivers a fully automated, AI‑driven coding assistant that integrates seamlessly into your GitHub codebase via the MCP protocol and a custom GitHub App. By labeling issues with remote-dev-ai, teams invoke our backend—hosted on AWS EC2 and AWS Lambda—that:

1. Fetches issue context
2. Generates code with Anthropic Claude
3. Opens a pull request with suggested changes

This model targets the 75% of developer time spent on non‑coding, repetitive tasks, slashes weekly issue‑resolution costs by 70% (from $1,200 in dev hours to $350 in AI calls), and provides a globally scalable solution for enterprises and startups alike.

---

Project Technology Stack

Integration & Hosting

• GitHub App & MCP Bot built in Python, registered as a GitHub App.
• Hosted on AWS EC2 instances running our MCP server and client for persistent connections and low latency

Event Handling

• AWS Lambda functions triggered by GitHub webhooks
• Lambda invokes the MCP server to fetch issue data and pipeline it into the Backend AI engine

AI Code Generation

• Anthropic accessed via secured API keys

Front-end

• Vercel Platform

Project Impact and Importance

• Time Savings
  Developers currently waste ~75% of their workday on non‑coding tasks (documentation, debugging, manual workflows).
• Cost Efficiency
  At 100 issue calls/day, AI costs ≈$350/week vs. $1,200/week for a mid‑level developer at $48.65/hr—a 70% reduction in direct labor cost.
• Global Scalability
  As a GitHub‑native solution, Remote‑Dev‑AI deploys instantly across any public or private repo, serving organizations of all sizes worldwide.
• Democratizing Automation
  Empowers small teams and emerging markets with enterprise‑grade automation without large DevOps budgets or specialist hires.

---

Development Challenges Faced

1. Immature MCP Protocol

The GitHub MCP server was open‑sourced only weeks before the hackathon, with minimal documentation. Integrating it into a production‑grade app required deep protocol analysis and custom extensions.

2. Context Management

Ensuring AI‑generated code aligns with each repo’s conventions demanded a dynamic context loader that fetches README, lint rules, and recent commits before each generation.(IN PROGRESS)

TEST OUR APP:

• https://v0-remote-dev-ai.vercel.app/

DEMO VIDEO:

https://www.youtube.com/watch?v=Yo3f0XnsO30

TEAM: remote-dev-ai

• Yashwanth Alapati
• Rajath Reghunath
• Manan Prakashkumar Patel
• Ritvik Vasantha Kumar

REFERENCES:

• Our front end code by vercel: https://github.com/rrgodhorus/hofhacks-remote-dev-ai-home