vozonda 0.7.5

Turn sources into your podcast

Vozonda reads them, voices make waves, you listen.

What is Vozonda?

Give Vozonda articles, papers, PDFs, audio files, podcast episodes, screenshots, YouTube videos, RSS feeds or your own notes, and it turns them into the show you want. Audio sources are transcribed locally.

• Fourteen show formats: deep dive, quick brief, debate, critique, morning dispatch of your feeds, feature story, explainer, three-voice roundtable, and more.
• One to three voices, or a single narrator: in ten languages; translates sources into the language you want to hear.
• Watchlists: turn RSS feeds into a daily or weekly digest that fills your feed automatically.
• Every episode comes with: chapters, transcripts, takeaways, highlights, and shareable clips; read and fix the script before it is voiced.
• Self-hosted & private: runs on your hardware with Docker. Voice generation runs on CPU by default with Kokoro (no GPU required). No cloud logins, no telemetry, and no subscriptions.
• Personal RSS feed: publishes episodes to your private podcast feed with chapters and transcripts, compatible with any podcast player, and optionally to Nostr.

Which install do you want?

This package is the Python orchestrator only. It does not contain the web UI. You talk to it over its HTTP API (/docs lists the routes) or through the MCP server.

Installation

Needs Python 3.11+, ffmpeg on the PATH, and an LLM endpoint (a local Ollama or vLLM, or any OpenAI-compatible API).

pip install "vozonda[tts-kokoro]"       # orchestrator + Kokoro voices on CPU
# or: uv tool install "vozonda[tts-kokoro]"

Optional extras

Without a TTS extra the server starts, but it cannot voice an episode.

Usage

export VOZONDA_LLM_BASE=http://127.0.0.1:11434/v1   # your LLM endpoint
export VOZONDA_LLM_MODEL=qwen2.5:latest
vozonda                                            # or: vozonda-api

The API listens on http://127.0.0.1:8787 (VOZONDA_HOST, VOZONDA_PORT change it); open http://127.0.0.1:8787/docs for the routes. Every setting is in .env.example. Before you expose it beyond this computer, read the deployment guide: remote requests need the VOZONDA_TOKEN.

Docker Quickstart

The full app, with the web UI and all dependencies:

git clone https://github.com/Vozonda/vozonda.git
cd vozonda
docker compose up -d

Prebuilt multi-architecture containers are available from GitHub Container Registry:

docker pull ghcr.io/vozonda/vozonda:latest

Links & Community

• Website & Audio Samples: https://vozonda.com
• GitHub Repository: https://github.com/Vozonda/vozonda
• Bug Tracker: https://github.com/Vozonda/vozonda/issues
• License: MIT