podhoc.com is an AI-powered SaaS platform that transforms any content source into a ready-to-publish podcast.
Users paste a URL — a YouTube video, a PDF, a blog post, a document — and Podhoc fetches the transcript, generates a conversational podcast script using large language models, and produces the final audio with AI text-to-speech voices.
How it works
- Add a source — paste any URL (YouTube, PDF, DOCX, plain text, web pages)
- Transcript extraction — Podhoc fetches and caches the content automatically
- Script generation — AI providers (Google Gemini, OpenAI, Grok, DeepSeek) craft a natural podcast script from the source material
- Audio production — text-to-speech engines generate the final podcast audio
- Download & publish — get your podcast episode ready to distribute
Tech stack
Podhoc is built on AWS with a serverless architecture:
- Frontend: React 19 + TypeScript + Vite, served via CloudFront
- Backend: Python Lambda functions orchestrated by Step Functions
- Database: PostgreSQL on RDS
- Auth: AWS Cognito with social login (Google, Apple)
- Payments: Stripe integration with a credits-based model
- AI: Multi-provider architecture supporting Gemini, OpenAI, Grok, and DeepSeek for both script generation and text-to-speech
- Storage: S3 for transcripts, scripts, and audio files
The platform is operated by Nexo Apex SL from Barcelona, Spain.