Your vibe coding session, as a cinematic narrated movie. Not a screen recording — a story.
git clone https://github.com/pooriaarab/vibemovieResumen de Tools
# VibeMovie
**Your vibe coding session, as a movie.**
Transforms your AI-assisted coding activity into a cinematic, narrated video — complete with chapters, a storyline, dramatic music, and visual effects. Not a screen recording. A movie.
## Demo
[▶ Watch the launch video](branding/launch-video.mp4) — what if claude made a movie about your PR?
https://github.com/pooriaarab/vibemovie/raw/main/branding/launch-video.mp4
## Install
```bash
npm i -g vibemovie
# or run it directly
npx vibemovie --help
```
## Quick start
```bash
# render a session recap from a JSON events file (or stdin)
vibemovie render session.json
# → writes ./vibe-recap.html — open it in any browser
vibemovie render session.json --ratio 9:16 --template speedrun --out recap.html
cat session.json | vibemovie render
# expose the render tool to your agent
vibemovie mcp
```
The default engine is **Hyperframes**: the recap is a self-contained animated
HTML page rendered fully **offline** — **zero API keys**, no network, nothing
leaves your machine. (Gen-video providers like Sora/Wavespeed slot into the
cascade above it later; v0 ships the guaranteed floor.)
Input is a JSON array of events (or `{ "events": [...] }`):
```json
[
{ "kind": "task-done", "ts": 1769500000000, "agent": "claude-code", "cwd": "/repo/demo",
"payload": { "label": "Refactor auth middleware", "durationMin": 18 } },
{ "kind": "tests-pass", "ts": 1769503600000, "payload": { "passed": 42 } },
{ "kind": "pr-merged", "ts": 1769505100000, "payload": { "pr": 42, "branch": "main" } }
]
```
Library usage:
```ts
import { renderMovie } from 'vibemovie';
const { html, path } = await renderMovie(events, {
ratio: '16:9', // '16:9' | '9:16' | '1:1'
template: 'documentary', // 'documentary' | 'speedrun' | 'meme'
out: 'recap.html', // optional — omit to just get the HTML string
});
```
`buildScenes(events)` (events → scene list) and `renderHyperframes(scenes)` (scene list → HTML) are pure and exported for custom pipelines.
## Why Build This
- VibeReplay captures sessions — VibeMovie turns them into stories
- "I built a SaaS in 4 hours" is interesting — a cinematic version of it is unforgettable
- Content creators spend hours editing coding videos — this automates the entire pipeline
- Every VibeMovie shared is a viral ad for the tool
- Think: AI-generated documentary of your coding process
## Features
- **Story arc** — AI analyzes your session and creates a narrative: setup, struggle, breakthrough, deploy
- **Cinematic editing** — Automatic cuts, zooms, transitions, and focus on the interesting parts
- **AI narration** — Generated voiceover that explains the journey: "At minute 23, everything broke..."
- **Dynamic soundtrack** — Music that matches the emotional arc (powered by VibeRadio engine)
- **Code highlights** — Key code snippets rendered beautifully with syntax highlighting and annotations
- **Stats overlay** — Lines of code, files changed, tokens used, time elapsed
- **Chapters** — Auto-generated based on git commits and milestones
- **Multiple formats** — YouTube (16:9), X/TikTok (9:16), LinkedIn (1:1), GIF summary
- **Thumbnail generation** — AI-generated thumbnail optimized for clicks
- **Templates** — Documentary, tutorial, speedrun, meme
## Generation engine (cascade)
Video generation follows the shared `@vibe/core` cascade so it always works:
1. **Your existing model** — if your agent's provider does video (e.g. Sora via your
OpenAI), use it.
2. **A video-gen key you bring** — Wavespeed, Replicate, or similar.
3. **Hyperframes fallback** — pure HTML/CSS/JS animated "video" (a rendered,
animated web page). No generative-video model, no key, no network — always
available. (The prototype below _is_ a working Hyperframes recap.)
Options across all tiers: **sync/live** (scenes build as the agent works) or
**async** (rendered after a turn/session) — user-configurable per hook · **aspect
ratios** 16:9 / 9:16 / 1:1 / GIF · **sound on/off** (soundtrack via the VibeRadio
engine + optional narration) · **subtitles/captions** · **transitions & templates**
(documentary, tutorial, speedrun, meme) · **avatar narrator** — if you've set up a
personal likeness/voice (e.g. HeyGen, or an on-device avatar), the recap can be
narrated by your avatar instead of an abstract summary.
## Distribution
- **CLI** — `vibemovie render session.json` (today) · `--session <id>` / `--repo .` (planned)
- **npm package** — `npm install -g vibemovie`
- **MCP server** — `vibemovie mcp` exposes a `render` tool your agent can call
- **Claude Code skill** — `/vibemovie` to render your last session
- **skills.sh** — Listed on skills.sh marketplace
- **Web app** — Upload a git repo, get a movie of its entire history
## Tech Stack
- Node.js + TypeScript
- Remotion (video rendering)
- Claude API (narrative generation, story arc)
- ElevenLabs (voiceover)
- Tone.js (soundtrack — shared with VibeRadio)
- FFmpeg (post-processing)
## Prototype
Interactive, self-contained UX prototype (no build, no network): open
[`docs/prototype.html`](docs/prototype.html) in a browser — it plays a real
Hyperframes recap with a working scrubber, the cascade selector, and an avatar
narrator toggle.
## Local-first
Runs on your own machine. The Hyperframes tier renders fully offline; nothing leaves
your machine unless you opt into a hosted video model. Enforced by the `@vibe/core`
consent model.
## Vibe Suite
Part of the **Vibe Suite** — companion tools for agentic coding CLIs (Claude Code,
Codex, Cursor, Gemini, Grok, pi, Kimi, and other harnesses). Ships as **CLI + npm
package + MCP server**.
## Relationship to VibeReplay
- **VibeReplay** = raw capture + timelapse (quick, functional)
- **VibeMovie** = cinematic, narrated, story-driven (polished, shareable)
- VibeMovie can use VibeReplay recordings as input
## Revenue Potential
- Free tier: 1 movie/month, watermarked, 720p
- Pro: unlimited, 4K, custom branding, all templates
- Studio: team movies, custom narration voice, API access
Lo que la gente pregunta sobre vibemovie
¿Qué es pooriaarab/vibemovie?
+
pooriaarab/vibemovie es tools para el ecosistema de Claude AI. Your vibe coding session, as a cinematic narrated movie. Not a screen recording — a story. Tiene 0 estrellas en GitHub y se actualizó por última vez today.
¿Cómo se instala vibemovie?
+
Puedes instalar vibemovie clonando el repositorio (https://github.com/pooriaarab/vibemovie) o siguiendo las instrucciones del README en GitHub. ClaudeWave también te ofrece bloques de instalación rápida en esta misma página.
¿Es seguro usar pooriaarab/vibemovie?
+
pooriaarab/vibemovie aún no ha sido auditado por nuestro agente de seguridad. Revisa el repositorio original en GitHub antes de usarlo en producción.
¿Quién mantiene pooriaarab/vibemovie?
+
pooriaarab/vibemovie es mantenido por pooriaarab. La última actividad registrada en GitHub es de today, con 0 issues abiertos.
¿Hay alternativas a vibemovie?
+
Sí. En ClaudeWave puedes explorar tools similares en /categories/tools, ordenados por popularidad o actividad reciente.
Despliega vibemovie en tu cloud
Lleva este repo a producción en minutos. Cada plataforma genera su propio entorno con variables de entorno editables.
¿Mantienes este repo? Añade un badge a tu README
Pega el badge en tu README de GitHub para mostrar que está auditado por ClaudeWave. Cada badge enlaza de vuelta a esta página y muestra el Trust Score actual.
[](https://claudewave.com/repo/pooriaarab-vibemovie)<a href="https://claudewave.com/repo/pooriaarab-vibemovie"><img src="https://claudewave.com/api/badge/pooriaarab-vibemovie" alt="Featured on ClaudeWave: pooriaarab/vibemovie" width="320" height="64" /></a>Más Tools
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
An AI SKILL that provide design intelligence for building professional UI/UX multiple platforms
🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman
✨ Light and Fast AI Assistant. Support: Web | iOS | MacOS | Android | Linux | Windows
CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies
AI agent skill that researches any topic across Reddit, X, YouTube, HN, Polymarket, and the web - then synthesizes a grounded summary