The Reading Room · Vol. 01
Long-form takes on the models, the workflows, and the production craft we're learning while we build.
Three frontier video models. Three philosophies. We ran the same prompts through each and built the only comparison chart that matters for working creators.
Read the field guideSoft-launched by ByteDance and walked away with two of 2026's biggest crowns — best motion fidelity and best price-per-second at quality tier. Here's why it belongs on your roster.
Read the spotlightThe model that quietly raised the resolution ceiling on AI video. A practical guide to multi-shot narrative work, character anchoring and the prompts that make Kling sing.
Read the workflowFilm yourself once. Apply that performance to any character, in any style. The most underrated mode in modern AI video, with a step-by-step playbook for landing it.
Read the tutorialTwo images. One prompt. The most under-used pro feature in AI video — and the fastest way to get exactly the shot you're picturing.
Read the tutorialOn a one-off shot they're peers. On a five-shot scene with a single character, only one of them gets to the end with continuity intact.
Read the comparisonThe image model that finally gets text right, edits surgically, and ships native 4K. A working creator's guide to the most capable image engine on the iPhone.
Read the workflowMost prompts fail because they're written like search queries. The ones that win are written like shot descriptions. Here are the structures and patterns that hold up across every model.
Read the craft guideThe feature that finally solves the hardest problem in AI video — making the same character look like the same person across multiple shots. A guide for narrative work.
Read the tutorialA start-to-finish playbook for shooting an AI short film on your iPhone — concept, casting, blocking, generation, post, delivery. Six chapters, one device.
Read the playbook