Blog
Guides on AI workshops and media pipelines, and Lab notes from our own tests.
Guides

Our free AI orchestrator kit
Our free orchestrator kit on GitHub: one Claude Code session plans the work, hands it to helper agents, has every change reviewed and asks the owner little.
· 5 min read
A prompt is a brief: container, content and polish
Write better AI prompts for images, video and apps: name the container, describe the content, add the polish, then read the prompt literally.
· 2 min read
AI avatars for corporate video: three real examples
How companies use AI avatars for internal events, training and lectures, with lessons from projects for Vaisala, Hiab and Aalto University Studios.
· 2 min read
How to run an AI workshop for a marketing team
A practical outline of an AI workshop that works for mixed-skill marketing teams: what to cover, how to structure the day and why making things beats watching demos.
· 2 min read
Localising video with AI without losing the brand
How to adapt video, audio and copy for new markets with AI, keeping cultural accuracy and the brand intact, from work for Aalto, Hiab and Fremantle.
· 2 min read
What is an AI media pipeline?
An AI media pipeline turns one source asset into every version your channels, markets and audiences need. Here is what it is made of and when it is worth building one.
· 2 min read
Lab notes
Dated tests of AI tools. The tools are older; the method still holds.
Song Bird: turning sound into a 3D light structure
Song Bird is a browser app that draws music or a voice as a growing 3D light structure. How it works, what the four formations do, and how to try it.
Lab · Tested December 2025 · 3 min read
A short film in 48 hours: our Runway Gen:48 entry T.E.
Our 3:40 short film for Runway's 48-hour Gen:48 contest, April 2025: the tools, where the characters drifted, and what a marketing team can use.
Lab · Tested April 2025 · 3 min read
The first weeks with Sora: January 2025 tests
20 Sora clips from two days in January 2025: what storyboard mode did to a 5-second shot, the made-up text, and four habits to keep.
Lab · Tested January 2025 · 3 min read
ComfyUI and AnimateDiff: making a painting move in 2024
One painting, one open-source ComfyUI and AnimateDiff workflow, one 8-second loop from February 2024: what moved, what drifted, and what it costs.
Lab · Tested February 2024 · 3 min read
Runway Gen-2 tests: a rainforest in 4-second clips (2023)
32 four-second Runway Gen-2 clips of one rainforest story, summer 2023: strong mood, drifting faces, and the method that still holds.
Lab · Tested July 2023 · 3 min read