Twocast is an AI-powered podcast generator designed for creating bilingual, multi-language, two-person conversational podcasts. It offers an alternative to tools like NotebookLLM, enabling users to generate short, engaging podcasts from various content sources with a focus on ease of use and multi-language support.

How It Works

Twocast leverages multiple AI models for text-to-speech (TTS) and large language model (LLM) generation, supporting providers like Fish Audio, Minimax, and Google Gemini for TTS, and OpenRouter and x.ai for LLM tasks. Users can input topics, URLs, documents (PDF, DOC, TXT), or lists to generate podcast content, which includes audio, an outline, and a script. The system is designed to facilitate two-person conversations, simulating natural dialogue.

Quick Start & Requirements

• Local Setup: Requires Docker for textract-rest-api and ffmpeg-api. Also needs Node.js (via yarn) and a PostgreSQL database.
• Docker One-Click: Uses docker compose up.
• API Keys: Requires API keys for TTS services (Fish Audio, Minimax, or Google Gemini) and LLM services (OpenRouter, x.ai).
• Setup: Local setup involves multiple Docker containers and database configuration. Docker Compose simplifies deployment.

Highlighted Details

• Generates 3-5 minute podcasts with one-click.
• Supports content generation from topics, links, documents, and lists.
• Features multi-language capabilities and downloadable audio output.
• Integrates with Fish Audio, Minimax, and Google Gemini for TTS.

Maintenance & Community

• Community support is available via Discord and Telegram.
• The project acknowledges tailwind-nextjs-starter-blog-i18n.

Licensing & Compatibility

• The repository does not explicitly state a license in the provided README.

Limitations & Caveats

• The README advises prioritizing the "Local Setup" method if issues arise with the Docker one-click start.
• Detailed environment variable configuration is necessary for API integrations.