# How to Vibe Code Your Own Async (and Stop Paying for It)

> Make videos like you have a team

- Site: https://podcastle.ai
- Category: AI Video & Audio Editor
- Verdict: **Serious undertaking** (35/100 vibecodeable)
- Estimated effort: 6+ weeks of part-time work

## Verdict

Build a simplified personal clip-editing subset rather than trying to match a multi-million-dollar AI media platform.

Replicating an enterprise-grade AI video suite like Async requires coordinating heavy asynchronous video encoding, GPU inference pipelines, and a complex bi-directional text-to-timeline synchronization state machine. While an AI agent can quickly scaffold a Next.js frontend with basic API wrappers, building a robust, lag-free media editor with real-time transcription cutting and voice cloning will take weeks of hard debugging around browser media stream limits and background job queues.

### What you can't replicate

- Proprietary video generation models and massive GPU compute clusters
- Active ecosystem of community video templates with tens of thousands of remixes
- Enterprise-grade multi-user collaborative workspace infrastructure

## What it does

An all-in-one, AI-powered video creation and editing platform that allows users to record, generate videos from prompts, edit conversationally via text transcripts, clean audio, clone voices, and repurpose media into social clips.

### Core features

- AI video generation from text prompts and templates
- Text-based conversational video editing
- Audio enhancement and noise suppression
- Speech-to-text automated transcription
- Text-to-speech voiceovers and voice cloning
- Long-form video repurposing into short social clips
- Translation and dubbing with lip sync

## The business

### Pricing

- Free: $0/mo — Basic recording caps and limited transcription.
- Storyteller: $11.99/mo — For solo creators needing higher limits and AI enhancements.
- Pro: $23.99/mo — Advanced professional features and expanded limits.
- Business / Teams: $39.99/mo — Team collaboration workspaces and admin controls.

### Funding

$22M to $23.5M raised.
- Seed Round (2020): ~$1.75M
- Series A (August 2021): $7M
- Series A Extension / Growth Round (February 2024): $13.5M
Investors: Mosaic Ventures, RTP Global, Point Nine Capital, Sierra Ventures, AI Fund

Founded 2020.
Team size: 51-100.

## The hard parts

- Synchronizing a bi-directional text transcript state machine with a non-linear millisecond-precise video timeline
- Orchestrating heavy asynchronous ML pipelines for voice cloning, transcription, and generative video scenes
- Managing high-throughput cloud storage and media transcoding for 4K video streams
- Handling real-time browser recording and multi-track audio stream synchronization without drift

## How to vibe code Async

### Prerequisites

- Node.js (free): Runtime for building and running the full-stack TypeScript application.
- GitHub (free): Source code repository and deployment pipeline integration.
- OpenAI API Key (pay-as-you-go (~$10 starting credit)): Required for Whisper speech-to-text transcription and LLM processing.
- ElevenLabs API Key (free tier / $5/mo starter): Required for voice cloning and text-to-speech voiceovers.

### Recommended AI tools

- Claude Code: Best-in-class terminal agent for scaffolding complex full-stack web applications and debugging media APIs.
- Cursor: Ideal for iterative frontend UI work on the complex video timeline and chat components.

### Stack

- Frontend: Next.js with Tailwind CSS and shadcn/ui
- Backend: Next.js API routes / Server Actions with Trigger.dev for background jobs
- Database: Turso (SQLite at the edge) with Drizzle ORM
- Auth: better-auth
- Payments: none
- Other: OpenAI API (Whisper STT), ElevenLabs API (TTS & Voice cloning), Fal.ai API (Video generation), Cloudflare R2 (Media file storage)

### Hosting

- Vercel (Hosting the Next.js frontend and serverless API functions): $0/mo (Hobby tier)
- Fly.io (Running background worker tasks and heavy video processing queues):  ~$5/mo
- Cloudflare (R2 Object storage for user-uploaded raw videos and generated clips): $0-5/mo

### Build guide

1. **Project Scaffolding & Database Schema** — Initialize the Next.js application with TypeScript, Tailwind CSS, and Drizzle ORM connected to Turso. Set up user authentication using better-auth and define database schemas for projects, media assets, transcripts, and editing timelines.

```
Scaffold a new Next.js project with Tailwind CSS, TypeScript, and App Router. Set up Drizzle ORM configured with Turso SQLite. Implement better-auth with email/password authentication. Create database tables for 'projects' (id, userId, title, createdAt, updatedAt), 'assets' (id, projectId, url, type, duration), and 'transcripts' (id, projectId, text, jsonSegments). Create a clean dashboard UI using shadcn/ui components where users can create new projects and view their list of workspace media files.
```

2. **Media Upload & Cloudflare R2 Storage Pipeline** — Implement file upload functionality for raw audio and video assets, routing large binaries directly to Cloudflare R2 object storage with signed upload URLs.

```
Implement server actions in Next.js to generate pre-signed upload URLs for Cloudflare R2 object storage. Build a drag-and-drop file upload component in the frontend supporting MP4, MOV, and MP3 files. Once uploaded, save the asset metadata into the 'assets' table linked to the current project. Handle upload progress state and display video/audio previews using HTML5 media elements.
```

3. **Speech-to-Text Transcription Integration** — Integrate OpenAI's Whisper API to transcribe uploaded media files into time-stamped text segments stored as structured JSON.

```
Create a background job endpoint using Trigger.dev that takes an uploaded asset ID, downloads the media file from Cloudflare R2, and sends it to the OpenAI Whisper API. Parse the resulting transcription response to extract word-level or segment-level timestamps. Save this structured JSON data into the 'transcripts' table and expose an API route to fetch the transcript for any given project.
```

4. **Interactive Text-to-Timeline Editing Interface** — Build an interactive frontend transcript editor where clicking or deleting words updates a synchronized video timeline state.

```
Build a split-screen video editing workspace in React. The left panel displays the interactive text transcript with selectable word blocks; the right panel features a video player. Implement state management so clicking any word in the transcript jumps the video playhead to that exact millisecond timestamp. Allow users to select and delete text blocks, which updates a local editing timeline edittree JSON array representing trimmed video segments.
```

5. **AI Voice Cloning & Text-to-Speech Integration** — Integrate ElevenLabs API to allow users to generate synthetic voiceovers and clone voices directly within their project dashboard.

```
Integrate the ElevenLabs API into backend services. Build a voice management tab in the dashboard where users can upload a clean audio sample to clone a voice or select from pre-made voices. Create an input form where users can type a script, select a voice, and generate a new audio track. Save the resulting generated audio file to Cloudflare R2 and add it as a new audio asset layer to the active project timeline.
```

6. **Video Processing & Export Pipeline** — Implement an asynchronous background worker using Fly.io and FFmpeg to stitch together trimmed video segments, apply audio enhancements, and export the final rendered file.

```
Build a Node.js background worker service deployed on Fly.io that listens for render jobs. When a user clicks 'Export Video', send the project timeline edittree and asset references to this worker. Use FFmpeg installed in the container to slice, trim, and combine media tracks according to the edit decision list, then upload the rendered MP4 file back to Cloudflare R2 and notify the frontend via polling or WebSockets when export is complete.
```

### Cost vs paying

**Starting costs (one-time):**

- OpenAI API starting credit: $10
- ElevenLabs starter tier: $5
- Total: ~$15 one-time

**Ongoing costs (monthly):**

- Fly.io worker node: $5/mo
- Cloudflare R2 & Workers storage: $3/mo
- OpenAI Whisper & ElevenLabs usage: ~$5-10/mo
- Total: ~$13-18/mo

- Paying for the SaaS instead: $14.99/mo (Storyteller plan)
- Build time: 40-60 hours
- AI tool credits: $20 (Claude Pro / Cursor)
- Break-even: 1 month of heavy use (though paying for SaaS saves 50+ hours of engineering work)

## Sources

- [Async Official Website](https://async.com)
- [PitchBook - Async (Podcastle) Company Profile](https://pitchbook.com/profiles/company/453147-35)
- [Podnews - Podcastle Rebrands to Async](https://podnews.net/press/podcastle-rebrands-to-async)