MinusPod is a self-hosted server that removes ads before you ever hit play. It transcribes episodes with Whisper, uses an LLM to detect and cut ad segments, and gets smarter over time by building cross-episode ad patterns and learning from your corrections. Bring your own LLM -- Claude, Ollama, OpenRouter, or any OpenAI-compatible provider.
MinusPod is a self-hosted tool that detects and removes ads from podcast episodes. It works by transcribing episodes, using an LLM to identify ad segments from the transcript, then cutting them from the audio and serving a clean RSS feed you can subscribe to in any podcast app.
It handles dynamically inserted ads (DAI) using audio fingerprinting and builds a cross-episode pattern library that improves detection over time across episodes and podcast networks. When it gets something wrong, there’s a correction UI that feeds back into the system.
Runs in Docker, supports Ollama for fully local inference if you don’t want to use an external API.
Open-source podcast ad removal system that uses Whisper transcription and the Claude API (or local LLMs via Ollama) to automatically detect and strip host-read sponsorships, dynamically inserted ads, and pre/mid/post-roll breaks from podcast feeds. Serves modified RSS feeds compatible with any podcast app—features multi-pass detection, audio fingerprinting, and cross-episode pattern learning.
I have been using this tool daily for the last 3 months. In general, I feel that it works pretty well. Unfortunately, I don't have a real measurement of accuracy, but it feels like it works ~95% of the time.
Shameless plug but it will allow you to create on demand OpenVPN Endpoints on AWS. However you will be shifting trust to AWS. https://github.com/ttlequals0/autovpn