Feedhorn is a practical publication focused on free, open-source, and self-hosted AI tools.
The name comes from radio astronomy: a feed horn is the component at the focal point of a dish antenna that catches an incoming signal and passes it into the receiver. That is the job of this site — catch the useful signal about self-hosted AI, filter out the noise, and pass it on to you.
We publish guides, comparisons, workflow breakdowns, and deployment notes for people who want useful AI systems without depending entirely on expensive SaaS subscriptions.
Our goal is to explain practical trade-offs and help readers build systems they can inspect and maintain. Unless an article identifies an actual test environment and results, it should be read as an explanation based on its cited sources, not a hands-on benchmark.
AI tools assist with drafting and editing. Sources and examples still require review; AI output is not evidence. See our editorial policy for sourcing and correction practices.