Local-first AI tools for people who make things.
Red 11 Technologies builds AI software for creators and small teams. We start with the slowest parts of audio and video post-production, and we build models that run on your own machine.
Four tools, one idea: give people back the hours.
Raw podcast footage in. An edit-ready first draft out.
Speaker 100:01:12So the first thing we changed was the whole intake process, because nobody could find anything.
Speaker 200:01:19And that was the week everything moved over to the new system?
Speaker 100:01:24Right. We spelled every product name the way the client does, every time.
Accurate transcripts on your Mac, then writing from them.
Hours of footage, cut down to the best material.
Client messages in. Sorted tasks, questions and decisions out.
Built by working editors, for the work editors actually do.
On your machine first
Media stays on your computer. Many clients can’t upload footage to a third party, and they shouldn’t have to.
Our own models
Where general-purpose AI falls short on real production audio, we train small, specialized models of our own.
You stay in charge
Our tools propose; people decide. Every result opens in the software you already use.
Small audio models that run on a laptop.
Our first in-house model identifies who is speaking in a recorded conversation, entirely on-device. Speaker identification is the foundation for PodDraft and EasyScribe.
Read about the researchof words attributed to the correct speaker across about 26,700 words of two-person conversation, running fully on-device. Early internal benchmark.