AudioPod AI From $20/mo · Free plan 9.1 Visit UNESCO PDF-to-Podcast
github.com
What it is
UNESCO PDF-to-Podcast turns scientific PDFs into podcast-style conversations. Its pipeline extracts structured material with Docling, uses Ollama with Granite4 to draft a dialogue, then creates audio with VibeVoice. Scripts can target experts, the general public, young people, or novices, and can be generated in English, French, Chinese, Spanish, Arabic, or Russian. Included native voices are available for English and Chinese; for the other four languages, English voices are used unless users provide samples. Users can set the desired duration and choose two to four speakers; defaults are 10 minutes and two speakers. The open-source project is free under the MIT license and can run locally from the command line on Linux, macOS, Windows, or self-hosted setups. Installation requires Python 3.9 or later, Poetry, and Ollama. A GPU is recommended, though CPU use is possible. It currently accepts one PDF at a time, and lengthy documents may need to be split into chunks. Audio generation can demand substantial GPU memory, particularly with the 7B model.
Spec sheet
- Purpose
- The project converts scientific PDFs into podcast-style audio conversations using AI.Read 4 Oct 2026 · github.com
- Integrations
- The documented pipeline uses Docling for PDF extraction, Ollama with Granite4 for script generation, and VibeVoice for audio synthesis.Read 4 Oct 2026 · github.com
- Security
- The README describes a local Ollama-based script-generation setup and makes no specific security certification or compliance claim.Read 4 Oct 2026 · github.com
- Support
- The project points users to GitHub Issues, Discussions, and wiki documentation for support.Read 4 Oct 2026 · github.com
- License
- The repository’s LICENSE file identifies the project license as MIT.Read 4 Oct 2026 · github.com
- Maker
- The project’s pyproject.toml lists UNESCO Data & AI as its author contact.Read 4 Oct 2026 · github.com
- Audience options
- Scripts can be tailored for experts, the general public, young people, or novices.Read 4 Oct 2026 · github.com
- Audiences
- Scripts can be tailored for experts, the general public, young people, or novice audiences.Read 4 Oct 2026 · github.com
- Audio controls
- Users can set podcast duration and choose 2–4 speakers; the default duration is 10 minutes and default speaker count is 2.Read 4 Oct 2026 · github.com
- Hardware
- The README warns that audio generation requires significant GPU memory, especially with the optional 7B model.Read 4 Oct 2026 · github.com
- Install and run
- The maker documents installation with Poetry or pip in a Python environment and running the pipeline from the command line.Read 4 Oct 2026 · github.com
- Installation
- The project requires Python 3.9 or later, Poetry, and Ollama, and recommends a GPU while allowing a CPU fallback.Read 4 Oct 2026 · github.com
- Languages
- The project supports script generation in UNESCO’s six official languages: English, French, Chinese, Spanish, Arabic, and Russian.Read 4 Oct 2026 · github.com
- License and maturity
- The repository labels the project MIT License (TBD) and identifies version 0.2.0 as MVP Complete.Read 4 Oct 2026 · github.com
- Limits
- The project currently accepts a single PDF at a time, with batch processing described as planned; long PDFs may need chunking.Read 4 Oct 2026 · github.com
- Local setup
- The README provides installation instructions for macOS and Linux.Read 4 Oct 2026 · github.com
- Maker profile
- The UNESCO GitHub profile describes its code repository as an open-source hub for education, science, culture, and sustainable development.Read 4 Oct 2026 · github.com
- Models and tools
- The pipeline uses Docling for PDF extraction, Ollama with Granite4 for script generation, and VibeVoice for audio synthesis.Read 4 Oct 2026 · github.com
- Notable limits
- The README says the project currently handles one PDF at a time, long PDFs may need chunking, and audio generation needs significant GPU memory, especially with the 7B model.Read 4 Oct 2026 · github.com
- PDF processing
- Its pipeline extracts structured PDF content, generates a conversational script, and synthesizes podcast audio.Read 4 Oct 2026 · github.com
- Pipeline
- Its pipeline extracts PDF content, generates a conversational script, and synthesizes podcast audio.Read 4 Oct 2026 · github.com
- Podcast controls
- Users can set a target duration and select between two and four speakers.Read 4 Oct 2026 · github.com
- Requirements
- The README lists Python 3.9 or newer, Poetry, and Ollama as prerequisites, and recommends a GPU while allowing CPU fallback.Read 4 Oct 2026 · github.com
- Voice availability
- English and Chinese have included native voices; French, Spanish, Arabic, and Russian fall back to English voices unless users add their own samples.Read 4 Oct 2026 · github.com
- Voice support
- The README says English and Chinese have included native voices, while French, Spanish, Arabic, and Russian fall back to English voices unless users add voice samples.Read 4 Oct 2026 · github.com
Strong points and limits
Strong points
- Supports six script-generation languages.
- Offers four audience styles and two to four speakers.
- Lets users set target audio duration.
- Free and open source under the MIT license.
Limits
- Processes one PDF at a time.
- Long PDFs may need chunking.
- Audio generation can require significant GPU memory.
- Several languages use English voices without added samples.
Who it suits
It suits readers who want to turn scientific PDFs into audio dialogue for a chosen audience and language. Users should be comfortable with local setup and its Python, Poetry, and Ollama requirements.
Verdict
The project gives users control over audience, duration, and speaker count while keeping the pipeline local. Its single-document workflow and hardware demands are important constraints to weigh before setup.
Where Specifiction places it
- AI Podcast Generators#21 of 237.2
