Vibe Rails

An opinionated framework that keeps vibe coding from going off the rails

View on GitHub VS Code VS Code Extension
Or run the install command below
Windows
irm https://raw.githubusercontent.com/robstokes857/vibe-rails/main/Scripts/install.ps1 | iex
Linux
wget -qO- https://raw.githubusercontent.com/robstokes857/vibe-rails/main/Scripts/install.sh | bash
Mac
curl -fsSL https://raw.githubusercontent.com/robstokes857/vibe-rails/main/Scripts/install.sh | bash

Key Features

Sandboxed and Parallel Processing

Assign different LLMs to different tracks at the same time. Let Claude plan, Codex refactor, and Gemini test in parallel without stepping on each other's work.

  • Dedicated sandbox per lane so agents never overwrite each other
  • Parallel execution across different models and tasks
  • No cross-contamination between files, prompts, or branch state
  • Compare outputs and merge only the best result into main
3x
Concurrent model lanes
Faster
Shorter cycle time than serial AI runs
Safer
No agents overwriting each other
Parallel execution board live
Claude lane Architecture + planning Running
Codex lane Refactor implementation Running
Gemini lane Tests + validation Running
Result: three focused agents, one coordinated workflow, dramatically less waiting.

Environment Isolation, Like Conda for LLMs

Create purpose-built AI environments with custom settings, scoped permissions, and reproducible profiles for every team and workflow.

  • Custom settings per environment: prompts, args, and tool behavior
  • Grant special permissions only where needed, keep sensitive flows locked down
  • Separate config directories and identity per environment
  • Clone proven profiles and reuse them across projects
Environment profiles managed
production_safe Strict rules, limited commands, protected paths Safe
refactor_lane Expanded tool access for deep code transformations Balanced
research_fast High-speed experimentation in isolated sandbox Fast
Like Conda, but purpose-built for AI coding workflows and team-safe defaults.

Remote Terminal Access

Let your AI write code while you're away from your desk. Vibe Rails gives you a secure, real-time terminal session in your browser -- connected to your local machine through an encrypted WebSocket relay.

  • API key authentication -- only you can access your sessions
  • Monitor your LLM's progress from your phone, tablet, or any browser
  • Full interactive terminal -- step in and type commands at any time
  • Auto-synced terminal size so output renders perfectly
viberails -- remote session
$ claude "refactor the auth module"
Claude is working...
Editing src/auth/middleware.ts
Editing src/auth/tokens.ts
Running npm test -- --watch=false
All 47 tests passed.
---
You watched this happen from your phone.

More Reasons Teams Choose Vibe Rails

After isolation and remote control, these are the biggest performance multipliers.

Convention Over Configuration

A stable, opinionated vibe-code framework that replaces setup sprawl with predictable defaults.

  • Sensible conventions for environments, terminals, and guardrails.
  • Less decision fatigue and faster onboarding for new contributors.
  • Consistent workflows across projects, repos, and teams.

Why it matters: Spend time shipping code, not reconfiguring tooling.

Rule Enforcement

Set explicit standards and require the model to fix issues before code can move forward.

  • Define team rules for tests, complexity, naming, and reliability.
  • Apply WARN, COMMIT, or STOP enforcement by rule.
  • Block risky changes before they become review debt.

Why it matters: Quality gates run automatically, not after regressions.

Cross-LLM Learning

Capture what works once, then reuse it across providers instead of starting from scratch.

  • Reuse proven prompts between Claude, Codex, and Gemini.
  • Keep style and architecture decisions consistent when switching models.
  • Build team memory that survives vendor changes.

Why it matters: Your best workflows compound over time.

RAG Without The Rot

Inject relevant context at the right moment instead of dumping stale project history into every prompt.

  • Track repeated fixes, feature decisions, and file-level change patterns.
  • Surface only the context needed for the current task.
  • Reduce drift caused by outdated or noisy prompt baggage.

Why it matters: Higher-signal context produces cleaner code.

Few Shot Prompting

Teach your preferred coding style once and make every model follow the same playbook.

  • Turn high-quality examples into reusable prompt assets.
  • Make Gemini or Codex perform like your strongest baseline.
  • Keep output consistent across teams and projects.

Why it matters: Consistency improves without prompt micromanagement.

Token Savings

Spend tokens on what matters instead of repeatedly feeding the model your entire codebase.

  • Prioritize the most relevant files for each request.
  • Cut redundant context and reduce oversized prompts.
  • Lower cost while speeding up response cycles.

Why it matters: Better targeting delivers faster, cheaper iterations.

Multiple LLMs and Code Editors

Claude

Codex

Gemini

Copilot

VS Code

Built With Modern Tech

Backend

.NET 10.0 with Native AOT

Frontend

Vanilla JS + Bootstrap 5

Extension

TypeScript 5.9.3

Protocol

Model Context Protocol