v2.0now supporting 4 platforms

Describe it.
We'll build
the plan.

ViblyOS turns your project idea into structured task sequences, dependency graphs, and platform-optimized config files for AI coding assistants. No boilerplate. No guesswork.

  • 4 platforms: Claude Code, Cursor, Antigravity, Codex
  • AI-decomposed task graphs with dependencies
  • Context-aware prompts for each build phase
  • CLI with sync, watch mode, and replan
  • Free tier — 15 credits, no credit card

Free to start • 15 credits included • No credit card required

viblyos generateoutputlogs
$ viblyos generate --platforms claude_code,cursor
# Analyzing project description...
Parsed intent 23 requirements extracted
Clarifying questions 3 questions (skipped)
Decomposed 14 tasks, 3 phases, 47 prompts
Generated claude_code config
Generated cursor config
# Output tree:
├── claude-code/
├── CLAUDE.md ................. 4.2 KB
├── .claude/settings.json .. 1.1 KB
├── .claude/agents/ ........ 3 files
└── prompts/ ............... 8 files
└── cursor/
├── .cursor/rules .......... 3.8 KB
└── prompts/ ............... 8 files
Done in 52s. 15 credits → 15 remaining (pro)
$
Powered byAnthropic
|
4 platformsClaude Code, Cursor, Antigravity, Codex
|
Up to 100tasks per project
|
Up to 100prompts generated
One project. Four platforms.
Generate optimized configs for any combination. Add platforms via AI translation for 8 credits each.
Claude Code
Anthropic
Agentic AI assistant with tool use, sub-agents, and autonomous coding workflows.
CLAUDE.md4.2 KB
.claude/settings.json1.1 KB
.claude/agents/*.md3 files
prompts/*.md8 files
Cursor
NEW
AI-first code editor with agent mode, background tasks, and multi-file editing.
.cursor/rules3.8 KB
.cursor/agents/*.md3 files
prompts/*.md8 files
Antigravity
Google
AI-powered development platform with Gemini integration and workspace rules.
GEMINI.md3.9 KB
.antigravity/rules.md1.4 KB
prompts/*.md8 files
Codex
OpenAI
Autonomous coding agent with sandboxed execution and structured task handling.
AGENTS.md4.0 KB
codex.toml0.8 KB
prompts/*.md8 files
Real output. Not mockups.
Here's what ViblyOS actually generates for a task management app.
CLAUDE.md
# Project: TaskFlow
## Overview
A kanban-style task management app with
real-time collaboration and AI-powered
task decomposition.
## Tech Stack
- Next.js 16 + TypeScript strict
- Supabase (auth, DB, realtime)
- Tailwind CSS + shadcn/ui
- Vercel deployment
## Architecture Rules
- App Router only, no pages/
- Server components by default
- Client components only for interactivity
- All DB queries through server actions
## Current Phase: 1 of 3
Focus: Core data model + auth flow
prompts/phase-1-task-03.md
# Task 03: Implement Board CRUD
Phase 1 • Priority: high • Est: 45min
## Context
Auth (task-01) and database schema
(task-02) are complete. The boards table
exists with RLS policies.
## Requirements
1. Create server actions:
src/lib/actions/boards.ts
2. Implement getBoards, createBoard,
updateBoard, deleteBoard
3. Add optimistic updates in UI
## Depends On
task-01 (auth)
task-02 (schema)
## Verification
$ npm run typecheck && npm run lint
Five steps. Fully automated.
From plain English to platform-specific configs in 10-60 minutes depending on complexity.
>_

Parse Intent

Extract requirements, constraints, and tech preferences from your description.

AI-powered
?!

Clarify

Optionally ask 2-3 clarifying questions to fill in ambiguous details.

optional

Decompose

Break into tasks with dependencies, phases, priorities, and time estimates.

AI-powered
</>

Generate

Produce config files, agents, skills, and prompts for each selected platform.

parallel

Deliver

Download as ZIP, sync via CLI, or auto-regenerate prompts as you build.

done
Built for how developers actually work.
>_ CLI

CLI-First Workflow

Full CLI tool for syncing progress, replanning, and adding platforms — with more commands coming soon.

$ npm install -g viblyos
MULTI

Multi-Platform Generation

Generate for one platform, then add more with a single command. Supports all 4 platforms.

$ viblyos add_platform cursor
δ SYNC

State-Aware Sync

Watch mode detects local changes, compares to plan, and regenerates stale prompts. Zero-credit local analysis.

$ viblyos sync --watch
REPLAN

Replan on the Fly

Requirements changed? Replan with updated context. Completed tasks are preserved. History tracked.

$ viblyos replan "add auth"
CACHE

Prompt Caching

1-hour TTL ephemeral caching on all AI calls. Reduces latency and cost on retries and regenerations.

cache_hit: true \u2022 saved 12s
RETRY

Retry from Failed Step

If generation fails mid-pipeline, retry picks up from last_completed_step. No wasted credits or time.

resuming from step 3/5...
Up to 100
tasks per project
4
platforms
Up to 100
prompts generated
80%
less planning time
Simple. Credit-based. No surprises.
Start free. Upgrade when you need more.
free
$0 /month
15 credits per month. Perfect for trying it out.
  • 15 credits/month
  • All 4 platforms
  • Full CLI access
  • Download project bundles
  • Free prompt regeneration
  • ×Priority support
get_started_free
ActionCredits
Generate project (single platform)15
Each additional platform+8
Replan8
CLI sync3
Prompt regenerationfree
Local watch modefree
 ╔══════════════════════════════════════╗
 ║     ready to build something?       ║
 ╚══════════════════════════════════════╝

Stop planning. Start building.

Free to start. 15 credits. No credit card. Cancel anytime.